Analysis of Public scRNAseq data from Ankylosing Spondylitis and Healthy Samples
Connor Eck
2025-02-03
library(tibble)Overview
This project aims to analyze publicly available scRNAseq data by comparing 23 samples with ankylosing spondylitis to 21 healthy samples, all from blood. 2 healthy and 2 AS samples were collected by the Kuhn lab, 10 AS and 29 healthy samples were obtained from public data collected by the University of California, San Francisco (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE194315), and 4 AS samples were collected from ArrayExpress (https://www.ebi.ac.uk/biostudies/arrayexpress/studies/E-MTAB-10948?query=E-MTAB-10948). Some collected samples were not blood and were dropped. A total of 23 diseased and 21 healthy samples were used for this analysis.
The samples collected by the Kuhn lab were provided as fastq files. These samples were mapped to a reference genome with starSolo to generate the count matrix, barcode, and gene list files for Seurat. All samples were integrated into a Seurat pipeline to create a Seurat object with all of the sample information. This Seurat object was used as input for the following DESeq2 pipeline.
An an all-vs-all comparison of differential gene expression was performed using DESeq2 and creating volcano plots from these results. Additionally, the gene lists from DESeq2 were used as input for Enrichr, which is a pathway analysis tool. Following this analysis, a comparison of healthy to diseased samples was performed on 5 unique cell types, Naive CD4, CD14 Monocytes, Memory CD4, CD8 T, and FCGR3A Monocytes. Cell clusters were identified using the FindMarkers function in Seurat. The barcodes for these cell types were subsetted and DESeq2 was used to perfom differential gene expression on these subsets. Volcano plots were generated from these results and the gene lists were used as input for Enrichr.
Downloading Data
This bash script was used to download samples from NCBI by taking a list of accession numbers as input. Once downloaded to my local machine, the files could be transferred to Alpine using Globus web file transfer.
#!/bin/bash
# Initialize a counter for the number of files downloaded
count=0
# Loop through the list of accession numbers
while read -r accession; do
echo "Downloading $accession..."
# Download the SRA file
prefetch $accession
# Change to the directory containing the .sra file
if [ -d "$accession" ]; then
cd "$accession" || { echo "Failed to change directory to $accession."; continue; }
# Run the fastq-dump command
fastq-dump --split-files --gzip --skip-technical --readids --read-filter pass --dumpbase --clip "${accession}.sra"
# Check if the FASTQ files were successfully created
if [ -f "${accession}_1.fastq.gz" ] && [ -f "${accession}_2.fastq.gz" ]; then
echo "$accession downloaded and processed successfully."
count=$((count + 1))
else
echo "Failed to generate FASTQ files for $accession."
fi
# Return to the parent directory
cd ..
else
echo "Directory $accession not found."
fi
done < accessions.txt
# Print the total number of files processed
echo "$count files downloaded and processed successfully."Quality Control
This script was used to perform quality control using fastp on the fastq files collected by the Kuhn lab. Fasp removes low quality reads, corrects low-quality bases and removes short reads or long poly-G tails.
# Create the output directory if it doesn't exist
mkdir -p "$output_dir"
# Loop through all the R1 fastq files in the input directory
for file1 in "$input_dir"/*_pass_1.fastq.gz; do
# Derive the base name (without _R1.fastq.gz)
base_name=$(basename "$file1" _pass_1.fastq.gz)
# Corresponding R2 file
file2="${input_dir}/${base_name}_pass_2.fastq.gz"
# Run fastp on paired-end files
fastp -i "$file1" -I "$file2" \
-o "$output_dir/${base_name}_pass_1_trimmed.fastq.gz" \
-O "$output_dir/${base_name}_pass_2_trimmed.fastq.gz"
doneIndexing
This script was used for indexing the reference genome pulled from NCBI (https://www.ncbi.nlm.nih.gov/datasets/taxonomy/9606/).
#!/bin/bash
# Request resources for the job
#SBATCH --nodes=1
#SBATCH --partition=amilan
#SBATCH --qos=normal
#SBATCH --time=24:00:00
#SBATCH --mem=30G
#SBATCH --ntasks=4
#SBATCH --account=amc-general
#SBATCH --job-name=jonsson_star
#SBATCH --output=Jonsson_map%J.log
#SBATCH --error=Jonsson_map%J.err
#SBATCH --mail-user=connor.eck@cuanschutz.edu
#SBATCH --mail-type=END
#Load Star
module load star/2.7.10b
#Run indexing command
STAR --runThreadN 4 --runMode genomeGenerate --genomeDir genome_index_output/ --genomeFastaFiles ./genome_index/genomic.fasta --sjdbGTFfile ./genome_index/genomic.gtf --sjdbOverhang 100Mapping
This script was used for mapping the fastq files collected by the Kuhn lab using starSolo. Cell barcodes are 16 bases long and unique molecular identifiers are 12 bases long.
#!/bin/bash
# Request resources for the job
#SBATCH --nodes=1
#SBATCH --partition=amilan
#SBATCH --qos=normal
#SBATCH --time=24:00:00
#SBATCH --mem=200G
#SBATCH --ntasks=10
#SBATCH --account=amc-general
#SBATCH --job-name=kuhn_starsolo
#SBATCH --output=Kuhn_solo_map%J.log
#SBATCH --error=Kuhn_solo_map%J.err
#SBATCH --mail-user=connor.eck@cuanschutz.edu
#SBATCH --mail-type=END
# Load STAR module
module load star/2.7.10b
# Directory containing your fastq files
input_dir="/scratch/alpine/ceck@xsede.org/Kuhn_Data/Kuhn_fastp_output_Glob/"
output_dir="/scratch/alpine/ceck@xsede.org/Kuhn_Data/Glob_mapping_output_solo/"
genome_dir="/scratch/alpine/ceck@xsede.org/Kuhn_Data/genome_index_output/"
# Create the output directory if it doesn't exist
mkdir -p "$output_dir"
# Number of threads to use
threads=10
# Loop through all the R1 fastq files in the input directory
for fileR1 in "$input_dir"/*_R1_trimmed.fastq.gz ; do
# Get the base name of the file (without the directory and _R1_trimmed.fastq.gz)
base_name=$(basename "$fileR1" _R1_trimmed.fastq.gz)
# Define the corresponding R2 file
fileR2="$input_dir/${base_name}_R2_trimmed.fastq.gz"
# Check if the corresponding R2 file exists
if [[ ! -f "$fileR2" ]]; then
echo "R2 file for $fileR1 not found, skipping..."
continue
fi
# Run STARsolo on each pair of fastq files (R1 and R2)
STAR --runThreadN "$threads" \
--genomeDir "$genome_dir" \
--limitBAMsortRAM 200000000000 \
--readFilesIn "$fileR1" "$fileR2" \
--readFilesCommand zcat \
--outFileNamePrefix "$output_dir/${base_name}_" \
--outSAMtype BAM SortedByCoordinate \
--soloType CB_UMI_Simple \
--soloCBstart 1 --soloCBlen 16 \
--soloUMIstart 17 --soloUMIlen 12 \
--soloFeatures Gene \
--soloBarcodeMate 1 \
--soloCBwhitelist None \
--clip5pNbases 28 5
# Check if the command was successful
if [[ $? -eq 0 ]]; then
echo "STARsolo successfully processed $fileR1 and $fileR2"
else
echo "Error processing $fileR1 and $fileR2 with STARsolo"
fi
doneCombining Samples
This script was used to combine the count matrix, barcode, and gene list files from each sample into a single Seurat object. Metadata for each sample is also added including sample ID, disease status, and batch.
library(Seurat)
library(ggplot2)
library(Matrix)
# # Read the data and create Seurat objects for Clonal Fates # #
# AS01PB
clonal_mat_AS01PB <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Clonal_fates/E-MTAB-10948/AS01PB.matrix.mtx.gz")
features_AS01PB <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Clonal_fates/E-MTAB-10948/AS01PB.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
barcodes_AS01PB <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Clonal_fates/E-MTAB-10948/AS01PB.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
features_AS01PB$V2 <- make.unique(features_AS01PB$V2)
rownames(clonal_mat_AS01PB) <- features_AS01PB$V2 # Gene names
colnames(clonal_mat_AS01PB) <- barcodes_AS01PB$V1 # Barcodes
AS01PB <- CreateSeuratObject(counts = clonal_mat_AS01PB, project = "Kuhn")
# Add sample identifiers, tissue type, cell type, and disease state
AS01PB$sample <- "AS01PB"
AS01PB$disease <- "ankylosing_spondylitis"
AS01PB$batch <- "clonal"
# AS01SF
clonal_mat_AS01SF <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Clonal_fates/E-MTAB-10948/AS01SF.matrix.mtx.gz")
features_AS01SF <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Clonal_fates/E-MTAB-10948/AS01SF.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
barcodes_AS01SF <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Clonal_fates/E-MTAB-10948/AS01SF.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
features_AS01SF$V2 <- make.unique(features_AS01SF$V2)
rownames(clonal_mat_AS01SF) <- features_AS01SF$V2 # Gene names
colnames(clonal_mat_AS01SF) <- barcodes_AS01SF$V1 # Barcodes
AS01SF <- CreateSeuratObject(counts = clonal_mat_AS01SF, project = "Kuhn")
# Add sample identifiers, tissue type, cell type, and disease state
AS01SF$sample <- "AS01SF"
AS01SF$cell_type <- "mononuclear_cell"
AS01SF$tissue_type <- "synovial_fluid"
AS01SF$disease <- "ankylosing_spondylitis"
AS01SF$batch <- "clonal"
# AS02PB
clonal_mat_AS02PB <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Clonal_fates/E-MTAB-10948/AS02PB.matrix.mtx.gz")
features_AS02PB <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Clonal_fates/E-MTAB-10948/AS02PB.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
barcodes_AS02PB <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Clonal_fates/E-MTAB-10948/AS02PB.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
features_AS02PB$V2 <- make.unique(features_AS02PB$V2)
rownames(clonal_mat_AS02PB) <- features_AS02PB$V2 # Gene names
colnames(clonal_mat_AS02PB) <- barcodes_AS02PB$V1 # Barcodes
AS02PB <- CreateSeuratObject(counts = clonal_mat_AS02PB, project = "Kuhn")
# Add sample identifiers, tissue type, cell type, and disease state
AS02PB$sample <- "AS02PB"
AS02PB$disease <- "ankylosing_spondylitis"
AS02PB$batch <- "clonal"
# AS02SF
clonal_mat_AS02SF <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Clonal_fates/E-MTAB-10948/AS02SF.matrix.mtx.gz")
features_AS02SF <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Clonal_fates/E-MTAB-10948/AS02SF.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
barcodes_AS02SF <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Clonal_fates/E-MTAB-10948/AS02SF.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
features_AS02SF$V2 <- make.unique(features_AS02SF$V2)
rownames(clonal_mat_AS02SF) <- features_AS02SF$V2 # Gene names
colnames(clonal_mat_AS02SF) <- barcodes_AS02SF$V1 # Barcodes
AS02SF <- CreateSeuratObject(counts = clonal_mat_AS02SF, project = "Kuhn")
# Add sample identifiers, tissue type, cell type, and disease state
AS02SF$sample <- "AS02SF"
AS02SF$cell_type <- "mononuclear_cell"
AS02SF$tissue_type <- "synovial_fluid"
AS02SF$disease <- "ankylosing_spondylitis"
AS02SF$batch <- "clonal"
# #Create Seurat Objects for Kuhn Matrices # #
#K_3_21B_1_S16_L001
mat_3_21B_1_S16_L001<- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Glob_mapping_output_solo/3_21B_1_S16_L001_Solo.out/Gene/filtered/matrix.mtx")
features_K_3_21B_1_S16_L001 <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Glob_mapping_output_solo/3_21B_1_S16_L001_Solo.out/Gene/filtered/features.tsv", header=FALSE, stringsAsFactors=FALSE)
barcodes_K_3_21B_1_S16_L001 <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Glob_mapping_output_solo/3_21B_1_S16_L001_Solo.out/Gene/filtered/barcodes.tsv", header=FALSE, stringsAsFactors=FALSE)
features_K_3_21B_1_S16_L001$V2 <- make.unique(features_K_3_21B_1_S16_L001$V2)
rownames(mat_3_21B_1_S16_L001) <- features_K_3_21B_1_S16_L001$V2 # Gene names
colnames(mat_3_21B_1_S16_L001) <- barcodes_K_3_21B_1_S16_L001$V1 # Barcodes
K_3_21B_1_S16_L001 <- CreateSeuratObject(counts = mat_3_21B_1_S16_L001, project = "Kuhn")
# Add sample identifiers, tissue type, cell type, and disease state
K_3_21B_1_S16_L001$sample <- "21B"
K_3_21B_1_S16_L001$disease <- "healthy"
K_3_21B_1_S16_L001$batch <- "Kuhn_collected"
#K_7_27B_1_S32_L001
mat_7_27B_1_S32_L001<- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Glob_mapping_output_solo/7_27B_1_S32_L001_Solo.out/Gene/filtered/matrix.mtx")
features_K_7_27B_1_S32_L001 <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Glob_mapping_output_solo/7_27B_1_S32_L001_Solo.out/Gene/filtered/features.tsv", header=FALSE, stringsAsFactors=FALSE)
barcodes_K_7_27B_1_S32_L001 <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Glob_mapping_output_solo/7_27B_1_S32_L001_Solo.out/Gene/filtered/barcodes.tsv", header=FALSE, stringsAsFactors=FALSE)
features_K_7_27B_1_S32_L001$V2 <- make.unique(features_K_7_27B_1_S32_L001$V2)
rownames(mat_7_27B_1_S32_L001) <- features_K_7_27B_1_S32_L001$V2 # Gene names
colnames(mat_7_27B_1_S32_L001) <- barcodes_K_7_27B_1_S32_L001$V1 # Barcodes
K_7_27B_1_S32_L001 <- CreateSeuratObject(counts = mat_7_27B_1_S32_L001, project = "Kuhn")
# Add sample identifiers, tissue type, cell type, and disease state
K_7_27B_1_S32_L001$sample <- "27B"
K_7_27B_1_S32_L001$disease <- "Spondyloarthritis"
K_7_27B_1_S32_L001$batch <- "Kuhn_collected"
# # Read the data and create Seurat objects for Epitope Paper # #
#PBMC_01_3
PBMC_01_3_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-01-3.matrix.mtx.gz")
PBMC_01_3_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-01-3.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_01_3_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-01-3.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_01_3_features$V2 <- make.unique(PBMC_01_3_features$V2)
rownames(PBMC_01_3_mat) <- PBMC_01_3_features$V2 # Gene names
colnames(PBMC_01_3_mat) <- PBMC_01_3_barcodes$V1 # Barcodes
PBMC_01_3 <- CreateSeuratObject(counts = PBMC_01_3_mat, project = "Kuhn")
PBMC_01_3$sample <- "PBMC_01_3_Healthy"
PBMC_01_3$disease <- "healthy"
PBMC_01_3$batch <- "epitope"
#PBMC_01_4
PBMC_01_4_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-01-4.matrix.mtx.gz")
PBMC_01_4_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-01-4.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_01_4_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-01-4.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_01_4_features$V2 <- make.unique(PBMC_01_4_features$V2)
rownames(PBMC_01_4_mat) <- PBMC_01_4_features$V2 # Gene names
colnames(PBMC_01_4_mat) <- PBMC_01_4_barcodes$V1 # Barcodes
PBMC_01_4 <- CreateSeuratObject(counts = PBMC_01_4_mat, project = "Kuhn")
PBMC_01_4$sample <- "PBMC_01_4_Healthy"
PBMC_01_4$disease <- "healthy"
PBMC_01_4$batch <- "epitope"
#PBMC_02_1
PBMC_02_1_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-02-1.matrix.mtx.gz")
PBMC_02_1_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-02-1.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_02_1_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-02-1.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
# Ensure unique feature names
PBMC_02_1_features$V2 <- make.unique(PBMC_02_1_features$V2)
# Assign row and column names to the matrix
rownames(PBMC_02_1_mat) <- PBMC_02_1_features$V2 # Gene names (features)
colnames(PBMC_02_1_mat) <- PBMC_02_1_barcodes$V1 # Barcodes (cells)
# Create the Seurat object
PBMC_02_1 <- CreateSeuratObject(counts = PBMC_02_1_mat, project = "Kuhn")
#Create AXI Object
PBMC_02_1_barcodes_AXI <- read.csv("PBMC_02_1_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_02_1_AXI <- subset(PBMC_02_1, cells = PBMC_02_1_barcodes_AXI)
PBMC_02_1_AXI$sample <- "PBMC_02_1_AXI"
PBMC_02_1_AXI$disease <- "AXI"
PBMC_02_1_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_02_1_barcodes_healthy <- read.csv("PBMC_02_1_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_02_1_healthy <- subset(PBMC_02_1, cells = PBMC_02_1_barcodes_healthy)
PBMC_02_1_healthy$sample <- "PBMC_02_1_Healthy"
PBMC_02_1_healthy$disease <- "healthy"
PBMC_02_1_healthy$batch <- "epitope"
#PBMC_02_2
PBMC_02_2_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-02-2.matrix.mtx.gz")
PBMC_02_2_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-02-2.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_02_2_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-02-2.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
# Ensure unique feature names
PBMC_02_2_features$V2 <- make.unique(PBMC_02_2_features$V2)
# Assign row and column names to the matrix
rownames(PBMC_02_2_mat) <- PBMC_02_2_features$V2 # Gene names (features)
colnames(PBMC_02_2_mat) <- PBMC_02_2_barcodes$V1 # Barcodes (cells)
# Create the Seurat object
PBMC_02_2 <- CreateSeuratObject(counts = PBMC_02_2_mat, project = "Kuhn")
#Create AXI Object
PBMC_02_2_barcodes_AXI <- read.csv("PBMC_02_2_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_02_2_AXI <- subset(PBMC_02_2, cells = PBMC_02_2_barcodes_AXI)
PBMC_02_2_AXI$sample <- "PBMC_02_2_AXI"
PBMC_02_2_AXI$disease <- "AXI"
PBMC_02_2_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_02_2_barcodes_healthy <- read.csv("PBMC_02_2_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_02_2_healthy <- subset(PBMC_02_2, cells = PBMC_02_2_barcodes_healthy)
PBMC_02_2_healthy$sample <- "PBMC_02_2_Healthy"
PBMC_02_2_healthy$disease <- "healthy"
PBMC_02_2_healthy$batch <- "epitope"
#PBMC_02_3
PBMC_02_3_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-02-3.matrix.mtx.gz")
PBMC_02_3_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-02-3.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_02_3_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-02-3.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
# Ensure unique feature names
PBMC_02_3_features$V2 <- make.unique(PBMC_02_3_features$V2)
# Assign row and column names to the matrix
rownames(PBMC_02_3_mat) <- PBMC_02_3_features$V2 # Gene names (features)
colnames(PBMC_02_3_mat) <- PBMC_02_3_barcodes$V1 # Barcodes (cells)
# Create the Seurat object
PBMC_02_3 <- CreateSeuratObject(counts = PBMC_02_3_mat, project = "Kuhn")
#Create AXI Object
PBMC_02_3_barcodes_AXI <- read.csv("PBMC_02_3_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_02_3_AXI <- subset(PBMC_02_3, cells = PBMC_02_3_barcodes_AXI)
PBMC_02_3_AXI$sample <- "PBMC_02_3_AXI"
PBMC_02_3_AXI$disease <- "AXI"
PBMC_02_3_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_02_3_barcodes_healthy <- read.csv("PBMC_02_3_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_02_3_healthy <- subset(PBMC_02_3, cells = PBMC_02_3_barcodes_healthy)
PBMC_02_3_healthy$sample <- "PBMC_02_3_Healthy"
PBMC_02_3_healthy$disease <- "healthy"
PBMC_02_3_healthy$batch <- "epitope"
#PBMC_02_4
PBMC_02_4_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-02-4.matrix.mtx.gz")
PBMC_02_4_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-02-4.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_02_4_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-02-4.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
# Ensure unique feature names
PBMC_02_4_features$V2 <- make.unique(PBMC_02_4_features$V2)
# Assign row and column names to the matrix
rownames(PBMC_02_4_mat) <- PBMC_02_4_features$V2 # Gene names (features)
colnames(PBMC_02_4_mat) <- PBMC_02_4_barcodes$V1 # Barcodes (cells)
# Create the Seurat object
PBMC_02_4 <- CreateSeuratObject(counts = PBMC_02_4_mat, project = "Kuhn")
#Create AXI Object
PBMC_02_4_barcodes_AXI <- read.csv("PBMC_02_4_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_02_4_AXI <- subset(PBMC_02_4, cells = PBMC_02_4_barcodes_AXI)
PBMC_02_4_AXI$sample <- "PBMC_02_4_AXI"
PBMC_02_4_AXI$disease <- "AXI"
PBMC_02_4_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_02_4_barcodes_healthy <- read.csv("PBMC_02_4_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_02_4_healthy <- subset(PBMC_02_4, cells = PBMC_02_4_barcodes_healthy)
PBMC_02_4_healthy$sample <- "PBMC_02_4_Healthy"
PBMC_02_4_healthy$disease <- "healthy"
PBMC_02_4_healthy$batch <- "epitope"
#PBMC_03_1
PBMC_03_1_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-03-1.matrix.mtx.gz")
PBMC_03_1_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-03-1.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_03_1_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-03-1.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
# Ensure unique feature names
PBMC_03_1_features$V2 <- make.unique(PBMC_03_1_features$V2)
# Assign row and column names to the matrix
rownames(PBMC_03_1_mat) <- PBMC_03_1_features$V2 # Gene names (features)
colnames(PBMC_03_1_mat) <- PBMC_03_1_barcodes$V1 # Barcodes (cells)
# Create the Seurat object
PBMC_03_1 <- CreateSeuratObject(counts = PBMC_03_1_mat, project = "Kuhn")
#Create AXI Object
PBMC_03_1_barcodes_AXI <- read.csv("PBMC_03_1_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_03_1_AXI <- subset(PBMC_03_1, cells = PBMC_03_1_barcodes_AXI)
PBMC_03_1_AXI$sample <- "PBMC_03_1_AXI"
PBMC_03_1_AXI$disease <- "AXI"
PBMC_03_1_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_03_1_barcodes_healthy <- read.csv("PBMC_03_1_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_03_1_healthy <- subset(PBMC_03_1, cells = PBMC_03_1_barcodes_healthy)
PBMC_03_1_healthy$sample <- "PBMC_03_1_Healthy"
PBMC_03_1_healthy$disease <- "healthy"
PBMC_03_1_healthy$batch <- "epitope"
#PBMC_03_2
PBMC_03_2_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-03-2.matrix.mtx.gz")
PBMC_03_2_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-03-2.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_03_2_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-03-2.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
# Ensure unique feature names
PBMC_03_2_features$V2 <- make.unique(PBMC_03_2_features$V2)
# Assign row and column names to the matrix
rownames(PBMC_03_2_mat) <- PBMC_03_2_features$V2 # Gene names (features)
colnames(PBMC_03_2_mat) <- PBMC_03_2_barcodes$V1 # Barcodes (cells)
# Create the Seurat object
PBMC_03_2 <- CreateSeuratObject(counts = PBMC_03_2_mat, project = "Kuhn")
#Create AXI Object
PBMC_03_2_barcodes_AXI <- read.csv("PBMC_03_2_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_03_2_AXI <- subset(PBMC_03_2, cells = PBMC_03_2_barcodes_AXI)
PBMC_03_2_AXI$sample <- "PBMC_03_2_AXI"
PBMC_03_2_AXI$disease <- "AXI"
PBMC_03_2_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_03_2_barcodes_healthy <- read.csv("PBMC_03_2_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_03_2_healthy <- subset(PBMC_03_2, cells = PBMC_03_2_barcodes_healthy)
PBMC_03_2_healthy$sample <- "PBMC_03_2_Healthy"
PBMC_03_2_healthy$disease <- "healthy"
PBMC_03_2_healthy$batch <- "epitope"
#PBMC_03_3
PBMC_03_3_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-03-3.matrix.mtx.gz")
PBMC_03_3_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-03-3.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_03_3_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-03-3.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
# Ensure unique feature names
PBMC_03_3_features$V2 <- make.unique(PBMC_03_3_features$V2)
# Assign row and column names to the matrix
rownames(PBMC_03_3_mat) <- PBMC_03_3_features$V2 # Gene names (features)
colnames(PBMC_03_3_mat) <- PBMC_03_3_barcodes$V1 # Barcodes (cells)
# Create the Seurat object
PBMC_03_3 <- CreateSeuratObject(counts = PBMC_03_3_mat, project = "Kuhn")
#Create AXI Object
PBMC_03_3_barcodes_AXI <- read.csv("PBMC_03_3_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_03_3_AXI <- subset(PBMC_03_3, cells = PBMC_03_3_barcodes_AXI)
PBMC_03_3_AXI$sample <- "PBMC_03_3_AXI"
PBMC_03_3_AXI$disease <- "AXI"
PBMC_03_3_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_03_3_barcodes_healthy <- read.csv("PBMC_03_3_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_03_3_healthy <- subset(PBMC_03_3, cells = PBMC_03_3_barcodes_healthy)
PBMC_03_3_healthy$sample <- "PBMC_03_3_Healthy"
PBMC_03_3_healthy$disease <- "healthy"
PBMC_03_3_healthy$batch <- "epitope"
#PBMC_03_4
PBMC_03_4_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-03-4.matrix.mtx.gz")
PBMC_03_4_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-03-4.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_03_4_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-03-4.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
# Ensure unique feature names
PBMC_03_4_features$V2 <- make.unique(PBMC_03_4_features$V2)
# Assign row and column names to the matrix
rownames(PBMC_03_4_mat) <- PBMC_03_4_features$V2 # Gene names (features)
colnames(PBMC_03_4_mat) <- PBMC_03_4_barcodes$V1 # Barcodes (cells)
# Create the Seurat object
PBMC_03_4 <- CreateSeuratObject(counts = PBMC_03_4_mat, project = "Kuhn")
#Create AXI Object
PBMC_03_4_barcodes_AXI <- read.csv("PBMC_03_4_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_03_4_AXI <- subset(PBMC_03_4, cells = PBMC_03_4_barcodes_AXI)
PBMC_03_4_AXI$sample <- "PBMC_03_4_AXI"
PBMC_03_4_AXI$disease <- "AXI"
PBMC_03_4_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_03_4_barcodes_healthy <- read.csv("PBMC_03_4_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_03_4_healthy <- subset(PBMC_03_4, cells = PBMC_03_4_barcodes_healthy)
PBMC_03_4_healthy$sample <- "PBMC_03_4_Healthy"
PBMC_03_4_healthy$disease <- "healthy"
PBMC_03_4_healthy$batch <- "epitope"
#PBMC_04_1
PBMC_04_1_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-04-1.matrix.mtx.gz")
PBMC_04_1_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-04-1.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_04_1_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-04-1.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_04_1_features$V2 <- make.unique(PBMC_04_1_features$V2)
rownames(PBMC_04_1_mat) <- PBMC_04_1_features$V2 # Gene names
colnames(PBMC_04_1_mat) <- PBMC_04_1_barcodes$V1 # Barcodes
PBMC_04_1 <- CreateSeuratObject(counts = PBMC_04_1_mat, project = "Kuhn")
#Create AXI Object
PBMC_04_1_barcodes_AXI <- read.csv("PBMC_04_1_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_04_1_AXI <- subset(PBMC_04_1, cells = PBMC_04_1_barcodes_AXI)
PBMC_04_1_AXI$sample <- "PBMC_04_1_AXI"
PBMC_04_1_AXI$disease <- "AXI"
PBMC_04_1_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_04_1_barcodes_healthy <- read.csv("PBMC_04_1_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_04_1_healthy <- subset(PBMC_04_1, cells = PBMC_04_1_barcodes_healthy)
PBMC_04_1_healthy$sample <- "PBMC_04_1_Healthy"
PBMC_04_1_healthy$disease <- "healthy"
PBMC_04_1_healthy$batch <- "epitope"
#PBMC_04_2
PBMC_04_2_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-04-2.matrix.mtx.gz")
PBMC_04_2_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-04-2.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_04_2_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-04-2.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_04_2_features$V2 <- make.unique(PBMC_04_2_features$V2)
rownames(PBMC_04_2_mat) <- PBMC_04_2_features$V2 # Gene names
colnames(PBMC_04_2_mat) <- PBMC_04_2_barcodes$V1 # Barcodes
PBMC_04_2 <- CreateSeuratObject(counts = PBMC_04_2_mat, project = "Kuhn")
#Create AXI Object
PBMC_04_2_barcodes_AXI <- read.csv("PBMC_04_2_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_04_2_AXI <- subset(PBMC_04_2, cells = PBMC_04_2_barcodes_AXI)
PBMC_04_2_AXI$sample <- "PBMC_04_2_AXI"
PBMC_04_2_AXI$disease <- "AXI"
PBMC_04_2_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_04_2_barcodes_healthy <- read.csv("PBMC_04_2_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_04_2_healthy <- subset(PBMC_04_2, cells = PBMC_04_2_barcodes_healthy)
PBMC_04_2_healthy$sample <- "PBMC_04_2_Healthy"
PBMC_04_2_healthy$disease <- "healthy"
PBMC_04_2_healthy$batch <- "epitope"
#PBMC_04_3
PBMC_04_3_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-04-3.matrix.mtx.gz")
PBMC_04_3_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-04-3.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_04_3_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-04-3.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_04_3_features$V2 <- make.unique(PBMC_04_3_features$V2)
rownames(PBMC_04_3_mat) <- PBMC_04_3_features$V2 # Gene names
colnames(PBMC_04_3_mat) <- PBMC_04_3_barcodes$V1 # Barcodes
PBMC_04_3 <- CreateSeuratObject(counts = PBMC_04_3_mat, project = "Kuhn")
#Create AXI Object
PBMC_04_3_barcodes_AXI <- read.csv("PBMC_04_3_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_04_3_AXI <- subset(PBMC_04_3, cells = PBMC_04_3_barcodes_AXI)
PBMC_04_3_AXI$sample <- "PBMC_04_3_AXI"
PBMC_04_3_AXI$disease <- "AXI"
PBMC_04_3_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_04_3_barcodes_healthy <- read.csv("PBMC_04_3_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_04_3_healthy <- subset(PBMC_04_3, cells = PBMC_04_3_barcodes_healthy)
PBMC_04_3_healthy$sample <- "PBMC_04_3_Healthy"
PBMC_04_3_healthy$disease <- "healthy"
PBMC_04_3_healthy$batch <- "epitope"
#PBMC_04_4
PBMC_04_4_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-04-4.matrix.mtx.gz")
PBMC_04_4_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-04-4.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_04_4_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-04-4.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_04_4_features$V2 <- make.unique(PBMC_04_4_features$V2)
rownames(PBMC_04_4_mat) <- PBMC_04_4_features$V2 # Gene names
colnames(PBMC_04_4_mat) <- PBMC_04_4_barcodes$V1 # Barcodes
PBMC_04_4 <- CreateSeuratObject(counts = PBMC_04_4_mat, project = "Kuhn")
#Create AXI Object
PBMC_04_4_barcodes_AXI <- read.csv("PBMC_04_4_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_04_4_AXI <- subset(PBMC_04_4, cells = PBMC_04_4_barcodes_AXI)
PBMC_04_4_AXI$sample <- "PBMC_04_4_AXI"
PBMC_04_4_AXI$disease <- "AXI"
PBMC_04_4_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_04_4_barcodes_healthy <- read.csv("PBMC_04_4_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_04_4_healthy <- subset(PBMC_04_4, cells = PBMC_04_4_barcodes_healthy)
PBMC_04_4_healthy$sample <- "PBMC_04_4_Healthy"
PBMC_04_4_healthy$disease <- "healthy"
PBMC_04_4_healthy$batch <- "epitope"
#PBMC_05_1
PBMC_05_1_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-05-1.matrix.mtx.gz")
PBMC_05_1_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-05-1.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_05_1_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-05-1.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_05_1_features$V2 <- make.unique(PBMC_05_1_features$V2)
rownames(PBMC_05_1_mat) <- PBMC_05_1_features$V2 # Gene names
colnames(PBMC_05_1_mat) <- PBMC_05_1_barcodes$V1 # Barcodes
PBMC_05_1 <- CreateSeuratObject(counts = PBMC_05_1_mat, project = "Kuhn")
#Create AXI Object
PBMC_05_1_barcodes_AXI <- read.csv("PBMC_05_1_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_05_1_AXI <- subset(PBMC_05_1, cells = PBMC_05_1_barcodes_AXI)
PBMC_05_1_AXI$sample <- "PBMC_05_1_AXI"
PBMC_05_1_AXI$disease <- "AXI"
PBMC_05_1_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_05_1_barcodes_healthy <- read.csv("PBMC_05_1_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_05_1_healthy <- subset(PBMC_05_1, cells = PBMC_05_1_barcodes_healthy)
PBMC_05_1_healthy$sample<- "PBMC_05_1_Healthy"
PBMC_05_1_healthy$disease <- "healthy"
PBMC_05_1_healthy$batch <- "epitope"
#PBMC_05_2
PBMC_05_2_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-05-2.matrix.mtx.gz")
PBMC_05_2_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-05-2.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_05_2_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-05-2.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_05_2_features$V2 <- make.unique(PBMC_05_2_features$V2)
rownames(PBMC_05_2_mat) <- PBMC_05_2_features$V2 # Gene names
colnames(PBMC_05_2_mat) <- PBMC_05_2_barcodes$V1 # Barcodes
PBMC_05_2 <- CreateSeuratObject(counts = PBMC_05_2_mat, project = "Kuhn")
#Create AXI Object
PBMC_05_2_barcodes_AXI <- read.csv("PBMC_05_2_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_05_2_AXI <- subset(PBMC_05_2, cells = PBMC_05_2_barcodes_AXI)
PBMC_05_2_AXI$sample <- "PBMC_05_2_AXI"
PBMC_05_2_AXI$disease <- "AXI"
PBMC_05_2_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_05_2_barcodes_healthy <- read.csv("PBMC_05_2_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_05_2_healthy <- subset(PBMC_05_2, cells = PBMC_05_2_barcodes_healthy)
PBMC_05_2_healthy$sample <- "PBMC_05_2_Healthy"
PBMC_05_2_healthy$disease <- "healthy"
PBMC_05_2_healthy$batch <- "epitope"
#PBMC_05_3
PBMC_05_3_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-05-3.matrix.mtx.gz")
PBMC_05_3_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-05-3.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_05_3_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-05-3.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_05_3_features$V2 <- make.unique(PBMC_05_3_features$V2)
rownames(PBMC_05_3_mat) <- PBMC_05_3_features$V2 # Gene names
colnames(PBMC_05_3_mat) <- PBMC_05_3_barcodes$V1 # Barcodes
PBMC_05_3 <- CreateSeuratObject(counts = PBMC_05_3_mat, project = "Kuhn")
#Create AXI Object
PBMC_05_3_barcodes_AXI <- read.csv("PBMC_05_3_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_05_3_AXI <- subset(PBMC_05_3, cells = PBMC_05_3_barcodes_AXI)
PBMC_05_3_AXI$sample <- "PBMC_05_3_AXI"
PBMC_05_3_AXI$disease <- "AXI"
PBMC_05_3_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_05_3_barcodes_healthy <- read.csv("PBMC_05_3_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_05_3_healthy <- subset(PBMC_05_3, cells = PBMC_05_3_barcodes_healthy)
PBMC_05_3_healthy$sample <- "PBMC_05_3_Healthy"
PBMC_05_3_healthy$disease <- "healthy"
PBMC_05_3_healthy$batch <- "epitope"
#PBMC_05_4
PBMC_05_4_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-05-4.matrix.mtx.gz")
PBMC_05_4_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-05-4.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_05_4_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-05-4.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_05_4_features$V2 <- make.unique(PBMC_05_4_features$V2)
rownames(PBMC_05_4_mat) <- PBMC_05_4_features$V2 # Gene names
colnames(PBMC_05_4_mat) <- PBMC_05_4_barcodes$V1 # Barcodes
PBMC_05_4 <- CreateSeuratObject(counts = PBMC_05_4_mat, project = "Kuhn")
#Create AXI Object
PBMC_05_4_barcodes_AXI <- read.csv("PBMC_05_4_bar.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_05_4_AXI <- subset(PBMC_05_4, cells = PBMC_05_4_barcodes_AXI)
PBMC_05_4_AXI$sample <- "PBMC_05_4_AXI"
PBMC_05_4_AXI$disease <- "AXI"
PBMC_05_4_AXI$batch <- "epitope"
# Create Healthy Object
PBMC_05_4_barcodes_healthy <- read.csv("PBMC_05_4_bar_Heal.csv", header = FALSE, stringsAsFactors = FALSE)$V1
PBMC_05_4_healthy <- subset(PBMC_05_4, cells = PBMC_05_4_barcodes_healthy)
PBMC_05_4_healthy$sample <- "PBMC_05_4_Healthy"
PBMC_05_4_healthy$disease <- "healthy"
PBMC_05_4_healthy$batch <- "epitope"
#PBMC_07_1
PBMC_07_1_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-07-1.matrix.mtx.gz")
PBMC_07_1_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-07-1.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_07_1_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-07-1.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_07_1_features$V2 <- make.unique(PBMC_07_1_features$V2)
rownames(PBMC_07_1_mat) <- PBMC_07_1_features$V2 # Gene names
colnames(PBMC_07_1_mat) <- PBMC_07_1_barcodes$V1 # Barcodes
PBMC_07_1 <- CreateSeuratObject(counts = PBMC_07_1_mat, project = "Kuhn")
# Add sample identifiers, tissue type, cell type, and disease state
PBMC_07_1$sample <- "PBMC_07_1"
PBMC_07_1$disease <- "AXI"
PBMC_07_1$batch <- "epitope"
#PBMC_07_2
PBMC_07_2_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-07-2.matrix.mtx.gz")
PBMC_07_2_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-07-2.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_07_2_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-07-2.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_07_2_features$V2 <- make.unique(PBMC_07_2_features$V2)
rownames(PBMC_07_2_mat) <- PBMC_07_2_features$V2 # Gene names
colnames(PBMC_07_2_mat) <- PBMC_07_2_barcodes$V1 # Barcodes
PBMC_07_2 <- CreateSeuratObject(counts = PBMC_07_2_mat, project = "Kuhn")
# Add sample identifiers, tissue type, cell type, and disease state
PBMC_07_2$sample <- "PBMC_07_2"
PBMC_07_2$disease <- "AXI"
PBMC_07_2$batch <- "epitope"
#PBMC_07_3
PBMC_07_3_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-07-3.matrix.mtx.gz")
PBMC_07_3_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-07-3.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_07_3_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-07-3.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_07_3_features$V2 <- make.unique(PBMC_07_3_features$V2)
rownames(PBMC_07_3_mat) <- PBMC_07_3_features$V2 # Gene names
colnames(PBMC_07_3_mat) <- PBMC_07_3_barcodes$V1 # Barcodes
PBMC_07_3 <- CreateSeuratObject(counts = PBMC_07_3_mat, project = "Kuhn")
# Add sample identifiers, tissue type, cell type, and disease state
PBMC_07_3$sample <- "PBMC_07_3"
PBMC_07_3$disease <- "AXI"
PBMC_07_3$batch <- "epitope"
#PBMC_07_4
PBMC_07_4_mat <- readMM("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-07-4.matrix.mtx.gz")
PBMC_07_4_features <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-07-4.features.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_07_4_barcodes <- read.table("/scratch/alpine/ceck@xsede.org/Kuhn_Data/Epitope_Count/GSE194315_PBMC-01-07_processed_data_files/PBMC-07-4.barcodes.tsv.gz", header=FALSE, stringsAsFactors=FALSE)
PBMC_07_4_features$V2 <- make.unique(PBMC_07_4_features$V2)
rownames(PBMC_07_4_mat) <- PBMC_07_4_features$V2 # Gene names
colnames(PBMC_07_4_mat) <- PBMC_07_4_barcodes$V1 # Barcodes
PBMC_07_4 <- CreateSeuratObject(counts = PBMC_07_4_mat, project = "Kuhn")
# Add sample identifiers, tissue type, cell type, and disease state
PBMC_07_4$sample <- "PBMC_07_4"
PBMC_07_4$disease <- "AXI"
PBMC_07_4$batch <- "epitope"
# List of Seurat objects
# List of Seurat objects
seurat_list <- list(
AS01PB,
AS02PB,
K_3_21B_1_S16_L001,
K_7_27B_1_S32_L001,
PBMC_02_1_AXI,
PBMC_02_2_AXI,
PBMC_02_3_AXI,
PBMC_02_4_AXI,
PBMC_03_1_AXI,
PBMC_03_2_AXI,
PBMC_03_3_AXI,
PBMC_03_4_AXI,
PBMC_04_1_AXI,
PBMC_04_2_AXI,
PBMC_04_3_AXI,
PBMC_04_4_AXI,
PBMC_05_1_AXI,
PBMC_05_2_AXI,
PBMC_05_3_AXI,
PBMC_05_4_AXI,
PBMC_01_3,
PBMC_01_4,
PBMC_02_1_healthy,
PBMC_02_2_healthy,
PBMC_02_3_healthy,
PBMC_02_4_healthy,
PBMC_03_1_healthy,
PBMC_03_2_healthy,
PBMC_03_3_healthy,
PBMC_03_4_healthy,
PBMC_04_1_healthy,
PBMC_04_2_healthy,
PBMC_04_3_healthy,
PBMC_04_4_healthy,
PBMC_05_1_healthy,
PBMC_05_2_healthy,
PBMC_05_3_healthy,
PBMC_05_4_healthy,
PBMC_07_1,
PBMC_07_2,
PBMC_07_3,
PBMC_07_4
)
# Extract unique features and integrate
#features <- SelectIntegrationFeatures(object.list = seurat_list)
#anchors <- FindIntegrationAnchors(object.list = seurat_list, anchor.features = features)
#merged_seurat <- IntegrateData(anchorset = anchors)
#saveRDS(merged_seurat, "merged_seurat_list.rds")
# Add back Unique features by adding raw data from original Seurat objects
#for (i in seq_along(seurat_list)) {
# Add back the original data with unique features from each object
# merged_seurat[['RNA']]@data <- cbind(merged_seurat[['RNA']]@data, seurat_list[[i]]@assays$RNA@data)
#}
merged_seurat<- merge(AS01PB, y = list(AS02PB,
K_3_21B_1_S16_L001,
K_7_27B_1_S32_L001,
PBMC_02_1_AXI,
PBMC_02_2_AXI,
PBMC_02_3_AXI,
PBMC_02_4_AXI,
PBMC_03_1_AXI,
PBMC_03_2_AXI,
PBMC_03_3_AXI,
PBMC_03_4_AXI,
PBMC_04_1_AXI,
PBMC_04_2_AXI,
PBMC_04_3_AXI,
PBMC_04_4_AXI,
PBMC_05_1_AXI,
PBMC_05_2_AXI,
PBMC_05_3_AXI,
PBMC_05_4_AXI,
PBMC_01_3,
PBMC_01_4,
PBMC_02_1_healthy,
PBMC_02_2_healthy,
PBMC_02_3_healthy,
PBMC_02_4_healthy,
PBMC_03_1_healthy,
PBMC_03_2_healthy,
PBMC_03_3_healthy,
PBMC_03_4_healthy,
PBMC_04_1_healthy,
PBMC_04_2_healthy,
PBMC_04_3_healthy,
PBMC_04_4_healthy,
PBMC_05_1_healthy,
PBMC_05_2_healthy,
PBMC_05_3_healthy,
PBMC_05_4_healthy,
PBMC_07_1,
PBMC_07_2,
PBMC_07_3,
PBMC_07_4
),
add.cell.ids = c("AS01PB",
"AS02PB",
"K_3_21B_1_S16_L001",
"K_7_27B_1_S32_L001",
"PBMC_02_1_AXI",
"PBMC_02_2_AXI",
"PBMC_02_3_AXI",
"PBMC_02_4_AXI",
"PBMC_03_1_AXI",
"PBMC_03_2_AXI",
"PBMC_03_3_AXI",
"PBMC_03_4_AXI",
"PBMC_04_1_AXI",
"PBMC_04_2_AXI",
"PBMC_04_3_AXI",
"PBMC_04_4_AXI",
"PBMC_05_1_AXI",
"PBMC_05_2_AXI",
"PBMC_05_3_AXI",
"PBMC_05_4_AXI",
"PBMC_01_3",
"PBMC_01_4",
"PBMC_02_1_healthy",
"PBMC_02_2_healthy",
"PBMC_02_3_healthy",
"PBMC_02_4_healthy",
"PBMC_03_1_healthy",
"PBMC_03_2_healthy",
"PBMC_03_3_healthy",
"PBMC_03_4_healthy",
"PBMC_04_1_healthy",
"PBMC_04_2_healthy",
"PBMC_04_3_healthy",
"PBMC_04_4_healthy",
"PBMC_05_1_healthy",
"PBMC_05_2_healthy",
"PBMC_05_3_healthy",
"PBMC_05_4_healthy",
"PBMC_07_1",
"PBMC_07_2",
"PBMC_07_3",
"PBMC_07_4"
),
project = "Kuhn")
saveRDS(merged_seurat, "merged_seurat_list.rds")
#features <- rownames(merged_seurat)
#write.table(features, file="features.txt", col.names = TRUE, row.names = FALSE, sep = "\t" )Batch Correction and Grouping Samples
This script was used to perform batch correction with harmony and to group the diseased samples. The diseases samples from separate sources did not have a standard naming system, so all of the diseased groups were assigned a new identifier called “AS_Group” and the healthy samples were assigned the “Healthy” identifier. UMAPs from before and after batch correction and sample grouping are shown below.
library(Seurat)
library(ggplot2)
library(harmony)
# Load the Seurat object
merged_seurat <- readRDS("merged_seurat_list.rds")
# Filter out NAs in the metadata if necessary
filtered_meta <- merged_seurat@meta.data[complete.cases(merged_seurat@meta.data), ]
merged_seurat <- subset(merged_seurat, cells = rownames(filtered_meta))
#Save features to a txt file
features <- rownames(merged_seurat)
write.table(features, file = "features.txt", col.names = TRUE, row.names = FALSE, sep = "\t")
# Create a new 'disease_grouped' column with only two groups: "AS_Group" and "Healthy"
merged_seurat@meta.data$disease_grouped <- ifelse(
merged_seurat@meta.data$disease %in% c("AXI", "ankylosing_spondylitis", "Spondyloarthritis"),
"AS_Group",
"Healthy"
)
# Visualize QC metrics as a violin plot
violin <- VlnPlot(merged_seurat , features = c("nFeature_RNA", "nCount_RNA"), ncol = 2)
#ggsave(filename = "violin_plot_K.pdf", plot = violin, width = 8, height = 6, dpi = 300)
feature_scatter <- FeatureScatter(merged_seurat , feature1 = "nCount_RNA", feature2 = "nFeature_RNA")
#ggsave(filename = "feature_scatter_plot_K.pdf", plot = feature_scatter, width = 8, height = 6, dpi = 300)
# Normalize the data
merged_seurat <- NormalizeData(merged_seurat)
# Identify highly variable features
merged_seurat <- FindVariableFeatures(merged_seurat, selection.method="vst", nfeatures=2000)
# Identify the 10 most highly variable genes
top10 <- head(VariableFeatures(merged_seurat ), 10)
# plot variable features with and without labels
var_no_label <- VariableFeaturePlot(merged_seurat)
var_label <- LabelPoints(plot = var_no_label, points = top10, repel = TRUE)
#ggsave(filename = "var_no_label_K.pdf", plot = var_no_label, width = 8, height = 6, dpi = 300)
#ggsave(filename = "var_label_K.pdf", plot = var_label, width = 8, height = 6, dpi = 300)
# Scale the data
all.genes <- rownames(merged_seurat)
merged_seurat <- ScaleData(merged_seurat, features = all.genes)
# Perform PCA
merged_seurat <- RunPCA(merged_seurat, features = VariableFeatures(object = merged_seurat))
loadings <- VizDimLoadings(merged_seurat, dims = 1:2, reduction = "pca")
#ggsave(filename = "loadings.pdf", plot = loadings, width = 8, height = 6, dpi = 300)
pca <- DimPlot(merged_seurat, reduction = "pca") + NoLegend()
#ggsave(filename = "pca.pdf", plot = pca, width = 8, height = 6, dpi = 300)
# Elbow plot
elbow <- ElbowPlot(merged_seurat)
#ggsave(filename = "elbow.pdf", plot = elbow, width = 8, height = 6, dpi = 300)
# Proceed with the Harmony and UMAP pipeline
merged_seurat <- RunHarmony(merged_seurat, group.by.vars = "sample")
merged_seurat <- FindNeighbors(merged_seurat, reduction = "harmony", dims = 1:10)
merged_seurat <- FindClusters(merged_seurat, resolution = 0.5)
merged_seurat <- RunUMAP(merged_seurat, reduction = "harmony", dims = 1:10)
saveRDS(merged_seurat, "umap_harmony.rds")
# Plot UMAP with the lumped disease categories
disease_lumped_plot <- DimPlot(merged_seurat, reduction = "umap", label = TRUE, group.by = "disease_grouped")
#ggsave(filename = "umap_disease_lumped_K.pdf", plot = disease_lumped_plot, width = 8, height = 6, dpi = 300)UMAP Before Batch correction and sample grouping
Before_Batch_correction.png
UMAP After Batch correction and sample grouping
After_Batch_correction.png
Dimensionality Reduction and Cluster Identification
This script was used to perform dimensionality reduction and cell cluster identification. The batch corrected Seurat object was used as input and the samples that were not healthy were still grouped into one group labeled AS_Group. Data normalization and PCA were performed. Integrated Seurat functions such as NormalizeData and RunUMAP were used for this step. The FindMarkers function was used for cell type identification and RunUMAP was used to generate a UMAP based on transcript counts.
library(Seurat)
library(ggplot2)
library(dplyr)
library(harmony)
library(presto)
# Load the Seurat object
merged_seurat <- readRDS("merged_seurat_list.rds")
# Filter out NAs in the metadata if necessary
filtered_meta <- merged_seurat@meta.data[complete.cases(merged_seurat@meta.data), ]
merged_seurat <- subset(merged_seurat, cells = rownames(filtered_meta))
#Save features to a txt file
features <- rownames(merged_seurat)
write.table(features, file = "features.txt", col.names = TRUE, row.names = FALSE, sep = "\t")
# Create a new 'disease_grouped' column with only two groups: "AS_Group" and "Healthy"
merged_seurat@meta.data$disease_grouped <- ifelse(
merged_seurat@meta.data$disease %in% c("AXI", "ankylosing_spondylitis", "Spondyloarthritis"),
"AS_Group",
"Healthy"
)
# Visualize QC metrics as a violin plot
violin <- VlnPlot(merged_seurat , features = c("nFeature_RNA", "nCount_RNA"), ncol = 2)
#ggsave(filename = "violin_plot_K.pdf", plot = violin, width = 8, height = 6, dpi = 300)
feature_scatter <- FeatureScatter(merged_seurat , feature1 = "nCount_RNA", feature2 = "nFeature_RNA")
#ggsave(filename = "feature_scatter_plot_K.pdf", plot = feature_scatter, width = 8, height = 6, dpi = 300)
# Normalize the data
merged_seurat <- NormalizeData(merged_seurat)
# Identify highly variable features
merged_seurat <- FindVariableFeatures(merged_seurat, selection.method="vst", nfeatures=2000)
# Identify the 10 most highly variable genes
top10 <- head(VariableFeatures(merged_seurat ), 10)
# plot variable features with and without labels
var_no_label <- VariableFeaturePlot(merged_seurat)
var_label <- LabelPoints(plot = var_no_label, points = top10, repel = TRUE)
#ggsave(filename = "var_no_label_K.pdf", plot = var_no_label, width = 8, height = 6, dpi = 300)
#ggsave(filename = "var_label_K.pdf", plot = var_label, width = 8, height = 6, dpi = 300)
# Scale the data
all.genes <- rownames(merged_seurat)
merged_seurat <- ScaleData(merged_seurat, features = all.genes)
# Perform PCA
merged_seurat <- RunPCA(merged_seurat, features = VariableFeatures(object = merged_seurat))
loadings <- VizDimLoadings(merged_seurat, dims = 1:2, reduction = "pca")
#ggsave(filename = "loadings.pdf", plot = loadings, width = 8, height = 6, dpi = 300)
pca <- DimPlot(merged_seurat, reduction = "pca") + NoLegend()
#ggsave(filename = "pca.pdf", plot = pca, width = 8, height = 6, dpi = 300)
# Elbow plot
elbow <- ElbowPlot(merged_seurat)
#ggsave(filename = "elbow.pdf", plot = elbow, width = 8, height = 6, dpi = 300)
# Proceed with the Harmony and UMAP pipeline
merged_seurat <- RunHarmony(merged_seurat, group.by.vars = "sample")
merged_seurat <- FindNeighbors(merged_seurat, reduction = "harmony", dims = 1:10)
merged_seurat <- FindClusters(merged_seurat, resolution = 0.5)
merged_seurat <- RunUMAP(merged_seurat, reduction = "harmony", dims = 1:10)
saveRDS(merged_seurat, "umap_harmony.rds")
# Plot UMAP with the lumped disease categories
disease_lumped_plot <- DimPlot(merged_seurat, reduction = "umap", label = TRUE, group.by = "disease_grouped")
#ggsave(filename = "umap_disease_lumped_K.pdf", plot = disease_lumped_plot, width = 8, height = 6, dpi = 300)
print(file.exists("/scratch/alpine/ceck@xsede.org/Kuhn_Data/barcode_filtering/new_pipe/NEWBATCH/lump/umap_harmony.rds"))
# Load the Seurat object
merged_seurat <- readRDS("/scratch/alpine/ceck@xsede.org/Kuhn_Data/barcode_filtering/new_pipe/NEWBATCH/lump/umap_harmony.rds")
# If needed, join layers (depending on your analysis setup)
merged_seurat <- JoinLayers(merged_seurat)
# Find markers for all clusters compared to the rest
merged_seurat.markers <- FindAllMarkers(merged_seurat, only.pos = TRUE)
# Extract the top 20 genes for each cluster
top_genes <- merged_seurat.markers %>%
group_by(cluster) %>%
dplyr::filter(avg_log2FC > 1) %>% # Filter for significant log fold changes
slice_max(order_by = avg_log2FC, n = 20) %>% # Get top 20 per cluster
ungroup()
# Rename clusters to more descriptive labels
new.cluster.ids <- c("Naive CD4 T", "CD14+ Mono", "Memory CD4 T", "B", "CD8 T",
"FCGR3A+ Mono", "NK", "DC", "Platelet", rep("Other", 20))
names(new.cluster.ids) <- levels(merged_seurat)
merged_seurat <- RenameIdents(merged_seurat, new.cluster.ids)
# Map numeric clusters to descriptive labels
#top_genes_output <- top_genes %>%
# mutate(cell_type = new.cluster.ids[as.numeric(cluster)]) %>% # Add cell type column
# select(gene, cluster, cell_type) # Select relevant columns
# Map numeric clusters to descriptive labels and include p-value columns
top_genes_output <- top_genes %>%
mutate(cell_type = new.cluster.ids[as.numeric(cluster)]) %>% # Add cell type column
select(gene, cluster, cell_type, avg_log2FC, p_val, p_val_adj)
# Save the top genes as a CSV file
write.csv(top_genes_output, "top20_genes_per_cluster_with_cell_types.csv", row.names = FALSE)
# Optional: Create a heatmap for visualization
top_genes_plot <- DoHeatmap(merged_seurat, features = top_genes$gene) + NoLegend()
ggsave(filename = "top_genes.pdf", plot = top_genes_plot, width = 20, height = 6, dpi = 300)
# Plot clusters with updated labels
cluster_ID <- DimPlot(merged_seurat, reduction = "umap", label = TRUE, pt.size = 0.5) + NoLegend()
ggsave(filename = "cluster_ID.pdf", plot = cluster_ID, width = 8, height = 6, dpi = 300)UMAP colored by Transcript counts
Transcript_Clusters.png
UMAP colored by cell type
Cell_Clusters.png
DEG analysis
This script was used to perform differential gene expression with DEseq2. Pseudobulking was executed based on disease grouping, cluster number, and sample ID. Groups were created using the Idents from the Seurat object. Diseased and healthy groups for each cell type were manually subsetted based on these Ident labels. Genes with counts of 0 were filtered out of the dataset and DESeq2 was run on the filtered object. Each comparison was saved as a rds object that contain gene lists, p-values, log fold change values, and adjusted p-values. For the final step, volcano plots were created using the log fold change and adjusted p-values from the DESeq2 results and the top 10 significant genes were labeled. Adjusted p-values were calculated using the Benjamini-Hochberg procedure to control the false discovery rate. Significance was defined as an adjusted p-value less than 0.05 and a log fold change greater than 1. Each plot was saved as a png.
library(Seurat)
library(ggplot2)
library(dplyr)
library(Matrix)
library(DESeq2)
library(ggrepel)
library(tidyr)
# Step 1: Load the Seurat Object
# Load the Seurat object
merged_seurat <- readRDS("/scratch/alpine/ceck@xsede.org/Kuhn_Data/umap_harmony.rds")
print("Data Loaded")
# Print the dimensions of the Seurat object
print(dim(merged_seurat))
# Check the cell distribution across combined groups
print(table(Idents(merged_seurat)))
# Step 2: Seperate the combined diseased groups and seperate healthy
# Create a new 'disease_grouped' column with two groups: "AS_Group" and "Healthy"
merged_seurat@meta.data$disease_grouped <- ifelse(
merged_seurat@meta.data$disease %in% c("AXI", "ankylosing_spondylitis", "Spondyloarthritis"),
"AS_Group",
"Healthy"
)
print("Groups Created")
# Join layers to access data
merged_seurat <- JoinLayers(merged_seurat)
# Step 3: Pseudobulk Data
# Aggregate expression data by cluster and treatment
pseudobulk_seurat <- AggregateExpression(
merged_seurat,
assays = "RNA",
group.by = c("disease_grouped", "seurat_clusters", "sample"),
return.seurat = TRUE
)
# Check the cell distribution across combined groups
print(table(Idents(pseudobulk_seurat)))
print("Pseudobulk Complete")
# Create a combined ID of disease group and cluster
pseudobulk_seurat$combined_id <- paste(
pseudobulk_seurat@meta.data$disease_grouped,
pseudobulk_seurat@meta.data$seurat_clusters,
pseudobulk_seurat@meta.data$sample,
sep = "_"
)
Idents(pseudobulk_seurat) <- "combined_id"
# Check the identities in pseudobulk_seurat
print("Available Idents:")
print(unique(Idents(pseudobulk_seurat)))
# Check the combined ID field
print("Combined IDs:")
print(pseudobulk_seurat$combined_id)
# Verify cluster names in the metadata
print("Unique clusters:")
print(unique(pseudobulk_seurat$seurat_clusters))
#Step 4: Subset the identities to be used in analysis
# Get the identities to keep
idents_to_keep <- c("AS-Group_0_27B", "AS-Group_0_AS01PB", "AS-Group_0_AS02PB", "AS-Group_0_PBMC-02-1-AXI",
"AS-Group_0_PBMC-02-2-AXI", "AS-Group_0_PBMC-02-3-AXI", "AS-Group_0_PBMC-02-4-AXI",
"AS-Group_0_PBMC-03-1-AXI", "AS-Group_0_PBMC-03-2-AXI", "AS-Group_0_PBMC-03-3-AXI",
"AS-Group_0_PBMC-03-4-AXI", "AS-Group_0_PBMC-04-1-AXI", "AS-Group_0_PBMC-04-2-AXI",
"AS-Group_0_PBMC-04-3-AXI", "AS-Group_0_PBMC-04-4-AXI", "AS-Group_0_PBMC-05-1-AXI",
"AS-Group_0_PBMC-05-2-AXI", "AS-Group_0_PBMC-05-3-AXI", "AS-Group_0_PBMC-05-4-AXI",
"AS-Group_0_PBMC-07-1", "AS-Group_0_PBMC-07-2", "AS-Group_0_PBMC-07-3", "AS-Group_0_PBMC-07-4",
"AS-Group_1_AS01PB", "AS-Group_1_AS02PB", "AS-Group_1_PBMC-02-1-AXI",
"AS-Group_1_PBMC-02-2-AXI", "AS-Group_1_PBMC-02-3-AXI", "AS-Group_1_PBMC-02-4-AXI",
"AS-Group_1_PBMC-03-1-AXI", "AS-Group_1_PBMC-03-2-AXI", "AS-Group_1_PBMC-03-3-AXI",
"AS-Group_1_PBMC-03-4-AXI", "AS-Group_1_PBMC-04-1-AXI", "AS-Group_1_PBMC-04-2-AXI",
"AS-Group_1_PBMC-04-3-AXI", "AS-Group_1_PBMC-04-4-AXI", "AS-Group_1_PBMC-05-1-AXI",
"AS-Group_1_PBMC-05-2-AXI", "AS-Group_1_PBMC-05-3-AXI", "AS-Group_1_PBMC-05-4-AXI",
"AS-Group_1_PBMC-07-1", "AS-Group_1_PBMC-07-2", "AS-Group_1_PBMC-07-3",
"AS-Group_1_PBMC-07-4",
"AS-Group_2_AS01PB", "AS-Group_2_AS02PB", "AS-Group_2_PBMC-02-1-AXI",
"AS-Group_2_PBMC-02-2-AXI", "AS-Group_2_PBMC-02-3-AXI", "AS-Group_2_PBMC-02-4-AXI",
"AS-Group_2_PBMC-03-1-AXI", "AS-Group_2_PBMC-03-2-AXI", "AS-Group_2_PBMC-03-3-AXI",
"AS-Group_2_PBMC-03-4-AXI", "AS-Group_2_PBMC-04-1-AXI", "AS-Group_2_PBMC-04-2-AXI",
"AS-Group_2_PBMC-04-3-AXI", "AS-Group_2_PBMC-04-4-AXI", "AS-Group_2_PBMC-05-1-AXI",
"AS-Group_2_PBMC-05-2-AXI", "AS-Group_2_PBMC-05-3-AXI", "AS-Group_2_PBMC-05-4-AXI",
"AS-Group_2_PBMC-07-1", "AS-Group_2_PBMC-07-2", "AS-Group_2_PBMC-07-3",
"AS-Group_2_PBMC-07-4",
"AS-Group_4_AS01PB", "AS-Group_4_AS02PB",
"AS-Group_4_PBMC-02-1-AXI", "AS-Group_4_PBMC-02-2-AXI",
"AS-Group_4_PBMC-02-3-AXI", "AS-Group_4_PBMC-02-4-AXI",
"AS-Group_4_PBMC-03-1-AXI", "AS-Group_4_PBMC-03-2-AXI",
"AS-Group_4_PBMC-03-3-AXI", "AS-Group_4_PBMC-03-4-AXI",
"AS-Group_4_PBMC-04-1-AXI", "AS-Group_4_PBMC-04-2-AXI",
"AS-Group_4_PBMC-04-3-AXI", "AS-Group_4_PBMC-04-4-AXI",
"AS-Group_4_PBMC-05-1-AXI", "AS-Group_4_PBMC-05-2-AXI",
"AS-Group_4_PBMC-05-3-AXI", "AS-Group_4_PBMC-05-4-AXI",
"AS-Group_4_PBMC-07-1", "AS-Group_4_PBMC-07-2",
"AS-Group_4_PBMC-07-3", "AS-Group_4_PBMC-07-4",
"AS-Group_5_AS01PB",
"AS-Group_5_AS02PB", "AS-Group_5_PBMC-02-1-AXI",
"AS-Group_5_PBMC-02-2-AXI", "AS-Group_5_PBMC-02-3-AXI",
"AS-Group_5_PBMC-02-4-AXI", "AS-Group_5_PBMC-03-1-AXI",
"AS-Group_5_PBMC-03-2-AXI", "AS-Group_5_PBMC-03-3-AXI",
"AS-Group_5_PBMC-03-4-AXI", "AS-Group_5_PBMC-04-1-AXI",
"AS-Group_5_PBMC-04-2-AXI", "AS-Group_5_PBMC-04-3-AXI",
"AS-Group_5_PBMC-04-4-AXI", "AS-Group_5_PBMC-05-1-AXI",
"AS-Group_5_PBMC-05-2-AXI", "AS-Group_5_PBMC-05-3-AXI",
"AS-Group_5_PBMC-05-4-AXI", "AS-Group_5_PBMC-07-1",
"AS-Group_5_PBMC-07-2", "AS-Group_5_PBMC-07-3",
"AS-Group_5_PBMC-07-4",
"Healthy_0_21B", "Healthy_0_PBMC-01-3-Healthy", "Healthy_0_PBMC-01-4-Healthy",
"Healthy_0_PBMC-02-1-Healthy", "Healthy_0_PBMC-02-2-Healthy", "Healthy_0_PBMC-02-3-Healthy",
"Healthy_0_PBMC-02-4-Healthy", "Healthy_0_PBMC-03-1-Healthy", "Healthy_0_PBMC-03-2-Healthy",
"Healthy_0_PBMC-03-3-Healthy", "Healthy_0_PBMC-03-4-Healthy", "Healthy_0_PBMC-04-1-Healthy",
"Healthy_0_PBMC-04-2-Healthy", "Healthy_0_PBMC-04-3-Healthy", "Healthy_0_PBMC-04-4-Healthy",
"Healthy_0_PBMC-05-1-Healthy", "Healthy_0_PBMC-05-2-Healthy", "Healthy_0_PBMC-05-3-Healthy",
"Healthy_0_PBMC-05-4-Healthy",
"Healthy_1_PBMC-01-3-Healthy", "Healthy_1_PBMC-01-4-Healthy",
"Healthy_1_PBMC-02-1-Healthy", "Healthy_1_PBMC-02-2-Healthy",
"Healthy_1_PBMC-02-3-Healthy", "Healthy_1_PBMC-02-4-Healthy",
"Healthy_1_PBMC-03-1-Healthy", "Healthy_1_PBMC-03-2-Healthy",
"Healthy_1_PBMC-03-3-Healthy", "Healthy_1_PBMC-03-4-Healthy",
"Healthy_1_PBMC-04-1-Healthy", "Healthy_1_PBMC-04-2-Healthy",
"Healthy_1_PBMC-04-3-Healthy", "Healthy_1_PBMC-04-4-Healthy",
"Healthy_1_PBMC-05-1-Healthy", "Healthy_1_PBMC-05-2-Healthy",
"Healthy_1_PBMC-05-3-Healthy", "Healthy_1_PBMC-05-4-Healthy",
"Healthy_2_PBMC-01-3-Healthy", "Healthy_2_PBMC-01-4-Healthy",
"Healthy_2_PBMC-02-1-Healthy", "Healthy_2_PBMC-02-2-Healthy",
"Healthy_2_PBMC-02-3-Healthy", "Healthy_2_PBMC-02-4-Healthy",
"Healthy_2_PBMC-03-1-Healthy", "Healthy_2_PBMC-03-2-Healthy",
"Healthy_2_PBMC-03-3-Healthy", "Healthy_2_PBMC-03-4-Healthy",
"Healthy_2_PBMC-04-1-Healthy", "Healthy_2_PBMC-04-2-Healthy",
"Healthy_2_PBMC-04-3-Healthy", "Healthy_2_PBMC-04-4-Healthy",
"Healthy_2_PBMC-05-1-Healthy", "Healthy_2_PBMC-05-2-Healthy",
"Healthy_2_PBMC-05-3-Healthy", "Healthy_2_PBMC-05-4-Healthy",
"Healthy_4_PBMC-01-3-Healthy",
"Healthy_4_PBMC-01-4-Healthy",
"Healthy_4_PBMC-02-1-Healthy", "Healthy_4_PBMC-02-2-Healthy", "Healthy_4_PBMC-02-3-Healthy",
"Healthy_4_PBMC-02-4-Healthy", "Healthy_4_PBMC-03-1-Healthy", "Healthy_4_PBMC-03-2-Healthy",
"Healthy_4_PBMC-03-3-Healthy", "Healthy_4_PBMC-03-4-Healthy", "Healthy_4_PBMC-04-1-Healthy",
"Healthy_4_PBMC-04-2-Healthy", "Healthy_4_PBMC-04-3-Healthy", "Healthy_4_PBMC-04-4-Healthy",
"Healthy_4_PBMC-05-1-Healthy", "Healthy_4_PBMC-05-2-Healthy", "Healthy_4_PBMC-05-3-Healthy",
"Healthy_4_PBMC-05-4-Healthy",
"Healthy_5_PBMC-01-3-Healthy", "Healthy_5_PBMC-01-4-Healthy",
"Healthy_5_PBMC-02-1-Healthy", "Healthy_5_PBMC-02-2-Healthy",
"Healthy_5_PBMC-02-3-Healthy", "Healthy_5_PBMC-02-4-Healthy",
"Healthy_5_PBMC-03-1-Healthy", "Healthy_5_PBMC-03-2-Healthy",
"Healthy_5_PBMC-03-3-Healthy", "Healthy_5_PBMC-03-4-Healthy",
"Healthy_5_PBMC-04-1-Healthy", "Healthy_5_PBMC-04-2-Healthy",
"Healthy_5_PBMC-04-3-Healthy", "Healthy_5_PBMC-04-4-Healthy",
"Healthy_5_PBMC-05-1-Healthy", "Healthy_5_PBMC-05-2-Healthy",
"Healthy_5_PBMC-05-3-Healthy", "Healthy_5_PBMC-05-4-Healthy"
)
# Subset the Seurat object
pseudobulk_seurat <- subset(pseudobulk_seurat, idents = idents_to_keep)
# Check the identities in the filtered object
print("Filtered Idents:")
table(Idents(pseudobulk_seurat))
# Check the identities in pseudobulk_seurat
#print(unique(Idents(pseudobulk_seurat)))
# Check the combined ID field
print("Combined IDs:")
print(pseudobulk_seurat$combined_id)
print("Seurat Object Subsetted Successfully")
# Step 5: Define the Comparisons for DEseq analysis from subsetted object
comparisons <- list(
AS_Group_vs_Healthy = list(
ident.1 = c("AS-Group_0_27B", "AS-Group_0_AS01PB", "AS-Group_0_AS02PB", "AS-Group_0_PBMC-02-1-AXI",
"AS-Group_0_PBMC-02-2-AXI", "AS-Group_0_PBMC-02-3-AXI", "AS-Group_0_PBMC-02-4-AXI",
"AS-Group_0_PBMC-03-1-AXI", "AS-Group_0_PBMC-03-2-AXI", "AS-Group_0_PBMC-03-3-AXI",
"AS-Group_0_PBMC-03-4-AXI", "AS-Group_0_PBMC-04-1-AXI", "AS-Group_0_PBMC-04-2-AXI",
"AS-Group_0_PBMC-04-3-AXI", "AS-Group_0_PBMC-04-4-AXI", "AS-Group_0_PBMC-05-1-AXI",
"AS-Group_0_PBMC-05-2-AXI", "AS-Group_0_PBMC-05-3-AXI", "AS-Group_0_PBMC-05-4-AXI",
"AS-Group_0_PBMC-07-1", "AS-Group_0_PBMC-07-2", "AS-Group_0_PBMC-07-3", "AS-Group_0_PBMC-07-4",
"AS-Group_1_AS01PB", "AS-Group_1_AS02PB", "AS-Group_1_PBMC-02-1-AXI",
"AS-Group_1_PBMC-02-2-AXI", "AS-Group_1_PBMC-02-3-AXI", "AS-Group_1_PBMC-02-4-AXI",
"AS-Group_1_PBMC-03-1-AXI", "AS-Group_1_PBMC-03-2-AXI", "AS-Group_1_PBMC-03-3-AXI",
"AS-Group_1_PBMC-03-4-AXI", "AS-Group_1_PBMC-04-1-AXI", "AS-Group_1_PBMC-04-2-AXI",
"AS-Group_1_PBMC-04-3-AXI", "AS-Group_1_PBMC-04-4-AXI", "AS-Group_1_PBMC-05-1-AXI",
"AS-Group_1_PBMC-05-2-AXI", "AS-Group_1_PBMC-05-3-AXI", "AS-Group_1_PBMC-05-4-AXI",
"AS-Group_1_PBMC-07-1", "AS-Group_1_PBMC-07-2", "AS-Group_1_PBMC-07-3",
"AS-Group_1_PBMC-07-4",
"AS-Group_2_AS01PB", "AS-Group_2_AS02PB", "AS-Group_2_PBMC-02-1-AXI",
"AS-Group_2_PBMC-02-2-AXI", "AS-Group_2_PBMC-02-3-AXI", "AS-Group_2_PBMC-02-4-AXI",
"AS-Group_2_PBMC-03-1-AXI", "AS-Group_2_PBMC-03-2-AXI", "AS-Group_2_PBMC-03-3-AXI",
"AS-Group_2_PBMC-03-4-AXI", "AS-Group_2_PBMC-04-1-AXI", "AS-Group_2_PBMC-04-2-AXI",
"AS-Group_2_PBMC-04-3-AXI", "AS-Group_2_PBMC-04-4-AXI", "AS-Group_2_PBMC-05-1-AXI",
"AS-Group_2_PBMC-05-2-AXI", "AS-Group_2_PBMC-05-3-AXI", "AS-Group_2_PBMC-05-4-AXI",
"AS-Group_2_PBMC-07-1", "AS-Group_2_PBMC-07-2", "AS-Group_2_PBMC-07-3",
"AS-Group_2_PBMC-07-4",
"AS-Group_4_AS01PB", "AS-Group_4_AS02PB",
"AS-Group_4_PBMC-02-1-AXI", "AS-Group_4_PBMC-02-2-AXI",
"AS-Group_4_PBMC-02-3-AXI", "AS-Group_4_PBMC-02-4-AXI",
"AS-Group_4_PBMC-03-1-AXI", "AS-Group_4_PBMC-03-2-AXI",
"AS-Group_4_PBMC-03-3-AXI", "AS-Group_4_PBMC-03-4-AXI",
"AS-Group_4_PBMC-04-1-AXI", "AS-Group_4_PBMC-04-2-AXI",
"AS-Group_4_PBMC-04-3-AXI", "AS-Group_4_PBMC-04-4-AXI",
"AS-Group_4_PBMC-05-1-AXI", "AS-Group_4_PBMC-05-2-AXI",
"AS-Group_4_PBMC-05-3-AXI", "AS-Group_4_PBMC-05-4-AXI",
"AS-Group_4_PBMC-07-1", "AS-Group_4_PBMC-07-2",
"AS-Group_4_PBMC-07-3", "AS-Group_4_PBMC-07-4",
"AS-Group_5_AS01PB",
"AS-Group_5_AS02PB", "AS-Group_5_PBMC-02-1-AXI",
"AS-Group_5_PBMC-02-2-AXI", "AS-Group_5_PBMC-02-3-AXI",
"AS-Group_5_PBMC-02-4-AXI", "AS-Group_5_PBMC-03-1-AXI",
"AS-Group_5_PBMC-03-2-AXI", "AS-Group_5_PBMC-03-3-AXI",
"AS-Group_5_PBMC-03-4-AXI", "AS-Group_5_PBMC-04-1-AXI",
"AS-Group_5_PBMC-04-2-AXI", "AS-Group_5_PBMC-04-3-AXI",
"AS-Group_5_PBMC-04-4-AXI", "AS-Group_5_PBMC-05-1-AXI",
"AS-Group_5_PBMC-05-2-AXI", "AS-Group_5_PBMC-05-3-AXI",
"AS-Group_5_PBMC-05-4-AXI", "AS-Group_5_PBMC-07-1",
"AS-Group_5_PBMC-07-2", "AS-Group_5_PBMC-07-3",
"AS-Group_5_PBMC-07-4"),
ident.2 = c("Healthy_0_21B", "Healthy_0_PBMC-01-3-Healthy", "Healthy_0_PBMC-01-4-Healthy",
"Healthy_0_PBMC-02-1-Healthy", "Healthy_0_PBMC-02-2-Healthy", "Healthy_0_PBMC-02-3-Healthy",
"Healthy_0_PBMC-02-4-Healthy", "Healthy_0_PBMC-03-1-Healthy", "Healthy_0_PBMC-03-2-Healthy",
"Healthy_0_PBMC-03-3-Healthy", "Healthy_0_PBMC-03-4-Healthy", "Healthy_0_PBMC-04-1-Healthy",
"Healthy_0_PBMC-04-2-Healthy", "Healthy_0_PBMC-04-3-Healthy", "Healthy_0_PBMC-04-4-Healthy",
"Healthy_0_PBMC-05-1-Healthy", "Healthy_0_PBMC-05-2-Healthy", "Healthy_0_PBMC-05-3-Healthy",
"Healthy_0_PBMC-05-4-Healthy",
"Healthy_1_PBMC-01-3-Healthy", "Healthy_1_PBMC-01-4-Healthy",
"Healthy_1_PBMC-02-1-Healthy", "Healthy_1_PBMC-02-2-Healthy",
"Healthy_1_PBMC-02-3-Healthy", "Healthy_1_PBMC-02-4-Healthy",
"Healthy_1_PBMC-03-1-Healthy", "Healthy_1_PBMC-03-2-Healthy",
"Healthy_1_PBMC-03-3-Healthy", "Healthy_1_PBMC-03-4-Healthy",
"Healthy_1_PBMC-04-1-Healthy", "Healthy_1_PBMC-04-2-Healthy",
"Healthy_1_PBMC-04-3-Healthy", "Healthy_1_PBMC-04-4-Healthy",
"Healthy_1_PBMC-05-1-Healthy", "Healthy_1_PBMC-05-2-Healthy",
"Healthy_1_PBMC-05-3-Healthy", "Healthy_1_PBMC-05-4-Healthy",
"Healthy_2_PBMC-01-3-Healthy", "Healthy_2_PBMC-01-4-Healthy",
"Healthy_2_PBMC-02-1-Healthy", "Healthy_2_PBMC-02-2-Healthy",
"Healthy_2_PBMC-02-3-Healthy", "Healthy_2_PBMC-02-4-Healthy",
"Healthy_2_PBMC-03-1-Healthy", "Healthy_2_PBMC-03-2-Healthy",
"Healthy_2_PBMC-03-3-Healthy", "Healthy_2_PBMC-03-4-Healthy",
"Healthy_2_PBMC-04-1-Healthy", "Healthy_2_PBMC-04-2-Healthy",
"Healthy_2_PBMC-04-3-Healthy", "Healthy_2_PBMC-04-4-Healthy",
"Healthy_2_PBMC-05-1-Healthy", "Healthy_2_PBMC-05-2-Healthy",
"Healthy_2_PBMC-05-3-Healthy", "Healthy_2_PBMC-05-4-Healthy",
"Healthy_4_PBMC-01-3-Healthy",
"Healthy_4_PBMC-01-4-Healthy",
"Healthy_4_PBMC-02-1-Healthy", "Healthy_4_PBMC-02-2-Healthy", "Healthy_4_PBMC-02-3-Healthy",
"Healthy_4_PBMC-02-4-Healthy", "Healthy_4_PBMC-03-1-Healthy", "Healthy_4_PBMC-03-2-Healthy",
"Healthy_4_PBMC-03-3-Healthy", "Healthy_4_PBMC-03-4-Healthy", "Healthy_4_PBMC-04-1-Healthy",
"Healthy_4_PBMC-04-2-Healthy", "Healthy_4_PBMC-04-3-Healthy", "Healthy_4_PBMC-04-4-Healthy",
"Healthy_4_PBMC-05-1-Healthy", "Healthy_4_PBMC-05-2-Healthy", "Healthy_4_PBMC-05-3-Healthy",
"Healthy_4_PBMC-05-4-Healthy",
"Healthy_5_PBMC-01-3-Healthy", "Healthy_5_PBMC-01-4-Healthy",
"Healthy_5_PBMC-02-1-Healthy", "Healthy_5_PBMC-02-2-Healthy",
"Healthy_5_PBMC-02-3-Healthy", "Healthy_5_PBMC-02-4-Healthy",
"Healthy_5_PBMC-03-1-Healthy", "Healthy_5_PBMC-03-2-Healthy",
"Healthy_5_PBMC-03-3-Healthy", "Healthy_5_PBMC-03-4-Healthy",
"Healthy_5_PBMC-04-1-Healthy", "Healthy_5_PBMC-04-2-Healthy",
"Healthy_5_PBMC-04-3-Healthy", "Healthy_5_PBMC-04-4-Healthy",
"Healthy_5_PBMC-05-1-Healthy", "Healthy_5_PBMC-05-2-Healthy",
"Healthy_5_PBMC-05-3-Healthy", "Healthy_5_PBMC-05-4-Healthy")
),
AS_Group_Naive_vs_Healthy_Naive = list(
ident.1 = c("AS-Group_0_27B", "AS-Group_0_AS01PB", "AS-Group_0_AS02PB", "AS-Group_0_PBMC-02-1-AXI",
"AS-Group_0_PBMC-02-2-AXI", "AS-Group_0_PBMC-02-3-AXI", "AS-Group_0_PBMC-02-4-AXI",
"AS-Group_0_PBMC-03-1-AXI", "AS-Group_0_PBMC-03-2-AXI", "AS-Group_0_PBMC-03-3-AXI",
"AS-Group_0_PBMC-03-4-AXI", "AS-Group_0_PBMC-04-1-AXI", "AS-Group_0_PBMC-04-2-AXI",
"AS-Group_0_PBMC-04-3-AXI", "AS-Group_0_PBMC-04-4-AXI", "AS-Group_0_PBMC-05-1-AXI",
"AS-Group_0_PBMC-05-2-AXI", "AS-Group_0_PBMC-05-3-AXI", "AS-Group_0_PBMC-05-4-AXI",
"AS-Group_0_PBMC-07-1", "AS-Group_0_PBMC-07-2", "AS-Group_0_PBMC-07-3", "AS-Group_0_PBMC-07-4"),
ident.2 = c("Healthy_0_21B", "Healthy_0_PBMC-01-3-Healthy", "Healthy_0_PBMC-01-4-Healthy",
"Healthy_0_PBMC-02-1-Healthy", "Healthy_0_PBMC-02-2-Healthy", "Healthy_0_PBMC-02-3-Healthy",
"Healthy_0_PBMC-02-4-Healthy", "Healthy_0_PBMC-03-1-Healthy", "Healthy_0_PBMC-03-2-Healthy",
"Healthy_0_PBMC-03-3-Healthy", "Healthy_0_PBMC-03-4-Healthy", "Healthy_0_PBMC-04-1-Healthy",
"Healthy_0_PBMC-04-2-Healthy", "Healthy_0_PBMC-04-3-Healthy", "Healthy_0_PBMC-04-4-Healthy",
"Healthy_0_PBMC-05-1-Healthy", "Healthy_0_PBMC-05-2-Healthy", "Healthy_0_PBMC-05-3-Healthy",
"Healthy_0_PBMC-05-4-Healthy")
),
AS_Group_CD14Mono_vs_Healthy_CD14Mono = list(
ident.1 = c("AS-Group_1_AS01PB", "AS-Group_1_AS02PB", "AS-Group_1_PBMC-02-1-AXI",
"AS-Group_1_PBMC-02-2-AXI", "AS-Group_1_PBMC-02-3-AXI", "AS-Group_1_PBMC-02-4-AXI",
"AS-Group_1_PBMC-03-1-AXI", "AS-Group_1_PBMC-03-2-AXI", "AS-Group_1_PBMC-03-3-AXI",
"AS-Group_1_PBMC-03-4-AXI", "AS-Group_1_PBMC-04-1-AXI", "AS-Group_1_PBMC-04-2-AXI",
"AS-Group_1_PBMC-04-3-AXI", "AS-Group_1_PBMC-04-4-AXI", "AS-Group_1_PBMC-05-1-AXI",
"AS-Group_1_PBMC-05-2-AXI", "AS-Group_1_PBMC-05-3-AXI", "AS-Group_1_PBMC-05-4-AXI",
"AS-Group_1_PBMC-07-1", "AS-Group_1_PBMC-07-2", "AS-Group_1_PBMC-07-3",
"AS-Group_1_PBMC-07-4"),
ident.2 = c("Healthy_1_PBMC-01-3-Healthy", "Healthy_1_PBMC-01-4-Healthy",
"Healthy_1_PBMC-02-1-Healthy", "Healthy_1_PBMC-02-2-Healthy",
"Healthy_1_PBMC-02-3-Healthy", "Healthy_1_PBMC-02-4-Healthy",
"Healthy_1_PBMC-03-1-Healthy", "Healthy_1_PBMC-03-2-Healthy",
"Healthy_1_PBMC-03-3-Healthy", "Healthy_1_PBMC-03-4-Healthy",
"Healthy_1_PBMC-04-1-Healthy", "Healthy_1_PBMC-04-2-Healthy",
"Healthy_1_PBMC-04-3-Healthy", "Healthy_1_PBMC-04-4-Healthy",
"Healthy_1_PBMC-05-1-Healthy", "Healthy_1_PBMC-05-2-Healthy",
"Healthy_1_PBMC-05-3-Healthy", "Healthy_1_PBMC-05-4-Healthy")
),
AS_Group_MemoryCD4_vs_Healthy_MemoryCD4 = list(
ident.1 = c("AS-Group_2_AS01PB", "AS-Group_2_AS02PB", "AS-Group_2_PBMC-02-1-AXI",
"AS-Group_2_PBMC-02-2-AXI", "AS-Group_2_PBMC-02-3-AXI", "AS-Group_2_PBMC-02-4-AXI",
"AS-Group_2_PBMC-03-1-AXI", "AS-Group_2_PBMC-03-2-AXI", "AS-Group_2_PBMC-03-3-AXI",
"AS-Group_2_PBMC-03-4-AXI", "AS-Group_2_PBMC-04-1-AXI", "AS-Group_2_PBMC-04-2-AXI",
"AS-Group_2_PBMC-04-3-AXI", "AS-Group_2_PBMC-04-4-AXI", "AS-Group_2_PBMC-05-1-AXI",
"AS-Group_2_PBMC-05-2-AXI", "AS-Group_2_PBMC-05-3-AXI", "AS-Group_2_PBMC-05-4-AXI",
"AS-Group_2_PBMC-07-1", "AS-Group_2_PBMC-07-2", "AS-Group_2_PBMC-07-3",
"AS-Group_2_PBMC-07-4"),
ident.2 = c("Healthy_2_PBMC-01-3-Healthy", "Healthy_2_PBMC-01-4-Healthy",
"Healthy_2_PBMC-02-1-Healthy", "Healthy_2_PBMC-02-2-Healthy",
"Healthy_2_PBMC-02-3-Healthy", "Healthy_2_PBMC-02-4-Healthy",
"Healthy_2_PBMC-03-1-Healthy", "Healthy_2_PBMC-03-2-Healthy",
"Healthy_2_PBMC-03-3-Healthy", "Healthy_2_PBMC-03-4-Healthy",
"Healthy_2_PBMC-04-1-Healthy", "Healthy_2_PBMC-04-2-Healthy",
"Healthy_2_PBMC-04-3-Healthy", "Healthy_2_PBMC-04-4-Healthy",
"Healthy_2_PBMC-05-1-Healthy", "Healthy_2_PBMC-05-2-Healthy",
"Healthy_2_PBMC-05-3-Healthy", "Healthy_2_PBMC-05-4-Healthy")
),
AS_Group_CD8T_vs_Healthy_CD8T = list(
ident.1 = c("AS-Group_4_AS01PB", "AS-Group_4_AS02PB",
"AS-Group_4_PBMC-02-1-AXI", "AS-Group_4_PBMC-02-2-AXI",
"AS-Group_4_PBMC-02-3-AXI", "AS-Group_4_PBMC-02-4-AXI",
"AS-Group_4_PBMC-03-1-AXI", "AS-Group_4_PBMC-03-2-AXI",
"AS-Group_4_PBMC-03-3-AXI", "AS-Group_4_PBMC-03-4-AXI",
"AS-Group_4_PBMC-04-1-AXI", "AS-Group_4_PBMC-04-2-AXI",
"AS-Group_4_PBMC-04-3-AXI", "AS-Group_4_PBMC-04-4-AXI",
"AS-Group_4_PBMC-05-1-AXI", "AS-Group_4_PBMC-05-2-AXI",
"AS-Group_4_PBMC-05-3-AXI", "AS-Group_4_PBMC-05-4-AXI",
"AS-Group_4_PBMC-07-1", "AS-Group_4_PBMC-07-2",
"AS-Group_4_PBMC-07-3", "AS-Group_4_PBMC-07-4"),
ident.2 = c("Healthy_4_PBMC-01-3-Healthy", "Healthy_4_PBMC-01-4-Healthy",
"Healthy_4_PBMC-02-1-Healthy", "Healthy_4_PBMC-02-2-Healthy", "Healthy_4_PBMC-02-3-Healthy",
"Healthy_4_PBMC-02-4-Healthy", "Healthy_4_PBMC-03-1-Healthy", "Healthy_4_PBMC-03-2-Healthy",
"Healthy_4_PBMC-03-3-Healthy", "Healthy_4_PBMC-03-4-Healthy", "Healthy_4_PBMC-04-1-Healthy",
"Healthy_4_PBMC-04-2-Healthy", "Healthy_4_PBMC-04-3-Healthy", "Healthy_4_PBMC-04-4-Healthy",
"Healthy_4_PBMC-05-1-Healthy", "Healthy_4_PBMC-05-2-Healthy", "Healthy_4_PBMC-05-3-Healthy",
"Healthy_4_PBMC-05-4-Healthy")
),
AS_Group_FCGR3AMono_vs_Healthy_FCGR3AMono = list(
ident.1 = c("AS-Group_5_AS01PB",
"AS-Group_5_AS02PB", "AS-Group_5_PBMC-02-1-AXI",
"AS-Group_5_PBMC-02-2-AXI", "AS-Group_5_PBMC-02-3-AXI",
"AS-Group_5_PBMC-02-4-AXI", "AS-Group_5_PBMC-03-1-AXI",
"AS-Group_5_PBMC-03-2-AXI", "AS-Group_5_PBMC-03-3-AXI",
"AS-Group_5_PBMC-03-4-AXI", "AS-Group_5_PBMC-04-1-AXI",
"AS-Group_5_PBMC-04-2-AXI", "AS-Group_5_PBMC-04-3-AXI",
"AS-Group_5_PBMC-04-4-AXI", "AS-Group_5_PBMC-05-1-AXI",
"AS-Group_5_PBMC-05-2-AXI", "AS-Group_5_PBMC-05-3-AXI",
"AS-Group_5_PBMC-05-4-AXI", "AS-Group_5_PBMC-07-1",
"AS-Group_5_PBMC-07-2", "AS-Group_5_PBMC-07-3",
"AS-Group_5_PBMC-07-4"),
ident.2 = c("Healthy_5_PBMC-01-3-Healthy", "Healthy_5_PBMC-01-4-Healthy",
"Healthy_5_PBMC-02-1-Healthy", "Healthy_5_PBMC-02-2-Healthy",
"Healthy_5_PBMC-02-3-Healthy", "Healthy_5_PBMC-02-4-Healthy",
"Healthy_5_PBMC-03-1-Healthy", "Healthy_5_PBMC-03-2-Healthy",
"Healthy_5_PBMC-03-3-Healthy", "Healthy_5_PBMC-03-4-Healthy",
"Healthy_5_PBMC-04-1-Healthy", "Healthy_5_PBMC-04-2-Healthy",
"Healthy_5_PBMC-04-3-Healthy", "Healthy_5_PBMC-04-4-Healthy",
"Healthy_5_PBMC-05-1-Healthy", "Healthy_5_PBMC-05-2-Healthy",
"Healthy_5_PBMC-05-3-Healthy", "Healthy_5_PBMC-05-4-Healthy")
)
)
print("Clusters Defined")
# Step 6: Filter out low expression genes
counts_matrix <- GetAssayData(pseudobulk_seurat, assay = "RNA", layer = "counts")
# Check the number of features before and after filtering
initial_features <- nrow(counts_matrix)
# Keep rows with elements greater than 1
#filtered_counts_matrix <- counts_matrix[rowSums(counts_matrix > 1) > 1, ]
# Keep rows with elements greater than 0
filtered_counts_matrix <- counts_matrix[rowSums(counts_matrix > 0) > 1, ]
filtered_features <- nrow(filtered_counts_matrix)
print(paste("Initial features:", initial_features))
print(paste("Filtered features:", filtered_features))
# Filter out low expression genes
#counts_matrix <- GetAssayData(pseudobulk_seurat, assay = "RNA", layer = "counts")
# Update the Seurat object with the filtered features
pseudobulk_seurat <- subset(pseudobulk_seurat, features = rownames(filtered_counts_matrix))
# Step 7: Perform Differential Expression Analysis
de_results_list <- lapply(names(comparisons), function(comp_name) {
comp <- comparisons[[comp_name]]
de_results <- FindMarkers(
object = pseudobulk_seurat,
ident.1 = comp$ident.1,
ident.2 = comp$ident.2,
test.use = "DESeq2"
)
de_results$gene <- rownames(de_results)
de_results$comparison <- comp_name
# Save each DE result as an .rds file
saveRDS(de_results, paste0("DESeq2_results_", comp_name, ".rds"))
return(de_results)
})
# Combine results into a single data frame
all_de_results <- do.call(rbind, de_results_list)
# Step 8: Annotate Data for Volcano Plots
all_de_results <- all_de_results %>%
mutate(
significance = case_when(
p_val_adj < 0.05 & abs(avg_log2FC) > 1 ~ "Significant",
TRUE ~ "Not Significant"
)
)
# Step 9: Generate Volcano Plots for Each Comparison
volcano_plots <- lapply(unique(all_de_results$comparison), function(comp_name) {
comp_data <- all_de_results %>% filter(comparison == comp_name)
# Identify top 10 significant genes based on p-value and log2FC
top_genes <- comp_data %>%
filter(significance == "Significant") %>%
arrange(p_val_adj) %>%
top_n(10, wt = -log10(p_val_adj)) %>%
pull(gene)
p <- ggplot(comp_data, aes(x = avg_log2FC, y = -log10(p_val_adj), color = significance)) +
geom_point(alpha = 0.8, size = 2) +
scale_color_manual(values = c("Significant" = "red", "Not Significant" = "gray")) +
theme_minimal() +
labs(
title = paste("Volcano Plot:", comp_name),
x = "Log2 Fold Change",
y = "-Log10 Adjusted P-Value"
) +
# Add labels for the top 10 significant genes
geom_text_repel(
data = comp_data %>% filter(gene %in% top_genes),
aes(label = gene),
box.padding = 0.35,
point.padding = 0.5,
max.overlaps = 10
)
# Save the plot
ggsave(paste0("volcano_plot_", comp_name, ".pdf"), plot = p, width = 8, height = 6)
return(p)
})
print("Pseudobulk analysis and volcano plots complete.")Results
This section contains the results for this project including volcano plots, gene lists, and Enrichr output.
Volcano Plots
Volcano Plots For Naive CD4 Cells
Naive Disease vs Healthy
Volcano Plots For Memory CD4 Cells
Memory CD4 Disease vs Healthy
Volcano Plots For CD8 T Cells
CD8 T Disease vs Healthy
Volcano Plots For FCGR3AMonocyte Cells
FCGR3A Monocyte Disease vs Healthy
Volcano Plots For CD14Monocyte Cells
CD14 Monocyte Disease vs Healthy
Volcano Plots For Healthy vs Control All Cells
Healthy vs Control All Cells
Gene Lists
Gene Lists For Naive CD4 Cells
data <- read.csv("Gene_Lists/DESeq2_results_AS_Group_Naive.csv")
knitr::kable(head(data, 10))| p_val | avg_log2FC | pct.1 | pct.2 | p_val_adj | comparison | status | gene |
|---|---|---|---|---|---|---|---|
| 0.0000018 | 3.0359866 | 0.826 | 0.947 | 0.0484896 | AS_Group_Naive_vs_Healthy_Naive | upregulated in diseased group | MTRNR2L12 |
| 0.0000284 | -0.4351223 | 0.870 | 0.947 | 0.7795041 | AS_Group_Naive_vs_Healthy_Naive | No significant fold change | FLYWCH1 |
| 0.0001354 | 0.0749019 | 0.696 | 0.895 | 1.0000000 | AS_Group_Naive_vs_Healthy_Naive | No significant fold change | C22orf34 |
| 0.0002778 | 0.2357586 | 0.957 | 0.947 | 1.0000000 | AS_Group_Naive_vs_Healthy_Naive | No significant fold change | AASDHPPT |
| 0.0003159 | 0.0978615 | 0.957 | 0.947 | 1.0000000 | AS_Group_Naive_vs_Healthy_Naive | No significant fold change | UQCC3 |
| 0.0003380 | 0.1988735 | 0.870 | 0.947 | 1.0000000 | AS_Group_Naive_vs_Healthy_Naive | No significant fold change | SUCLA2 |
| 0.0003712 | 0.1783328 | 0.565 | 0.947 | 1.0000000 | AS_Group_Naive_vs_Healthy_Naive | No significant fold change | RMRP |
| 0.0004321 | -0.3505566 | 0.957 | 0.947 | 1.0000000 | AS_Group_Naive_vs_Healthy_Naive | No significant fold change | GPS2 |
| 0.0007279 | 0.2258815 | 0.957 | 0.947 | 1.0000000 | AS_Group_Naive_vs_Healthy_Naive | No significant fold change | GLRX |
| 0.0007401 | 1.2106819 | 0.957 | 0.947 | 1.0000000 | AS_Group_Naive_vs_Healthy_Naive | upregulated in diseased group | MT-ATP8 |
Gene Lists For Memory CD4 Cells
data <- read.csv("Gene_Lists/DESeq2_results_AS_Group_MemoryCD4.csv")
knitr::kable(head(data, 10))| p_val | avg_log2FC | pct.1 | pct.2 | p_val_adj | comparison | status | gene |
|---|---|---|---|---|---|---|---|
| 0.0000016 | 3.0287072 | 0.818 | 1.000 | 0.0428071 | AS_Group_MemoryCD4_vs_Healthy_MemoryCD4 | upregulated in diseased group | MTRNR2L12 |
| 0.0000292 | -0.5246019 | 0.864 | 1.000 | 0.8016880 | AS_Group_MemoryCD4_vs_Healthy_MemoryCD4 | No significant fold change | ZBED6 |
| 0.0000794 | -0.3371374 | 0.864 | 1.000 | 1.0000000 | AS_Group_MemoryCD4_vs_Healthy_MemoryCD4 | No significant fold change | FLYWCH1 |
| 0.0001712 | 0.2130475 | 0.636 | 0.833 | 1.0000000 | AS_Group_MemoryCD4_vs_Healthy_MemoryCD4 | No significant fold change | TRAV4 |
| 0.0001720 | 0.6275713 | 0.818 | 1.000 | 1.0000000 | AS_Group_MemoryCD4_vs_Healthy_MemoryCD4 | No significant fold change | AC016831.7 |
| 0.0001787 | 0.1834658 | 1.000 | 1.000 | 1.0000000 | AS_Group_MemoryCD4_vs_Healthy_MemoryCD4 | No significant fold change | GTF3C6 |
| 0.0002454 | 0.3596492 | 0.818 | 1.000 | 1.0000000 | AS_Group_MemoryCD4_vs_Healthy_MemoryCD4 | No significant fold change | KIAA0391 |
| 0.0002533 | 1.3761931 | 0.955 | 1.000 | 1.0000000 | AS_Group_MemoryCD4_vs_Healthy_MemoryCD4 | upregulated in diseased group | MT-ATP8 |
| 0.0002574 | -0.6079710 | 0.955 | 1.000 | 1.0000000 | AS_Group_MemoryCD4_vs_Healthy_MemoryCD4 | No significant fold change | NBPF19 |
| 0.0003024 | 0.1786150 | 0.955 | 1.000 | 1.0000000 | AS_Group_MemoryCD4_vs_Healthy_MemoryCD4 | No significant fold change | DECR1 |
Gene Lists For CD8 T Cells
data <- read.csv("Gene_Lists/DESeq2_results_AS_Group_CD8T.csv")
knitr::kable(head(data, 10))| p_val | avg_log2FC | pct.1 | pct.2 | p_val_adj | comparison | status | gene |
|---|---|---|---|---|---|---|---|
| 0.0000544 | 0.2083657 | 0.636 | 0.889 | 1 | AS_Group_CD8T_vs_Healthy_CD8T | No significant fold change | AQP3 |
| 0.0000589 | 2.6149708 | 0.818 | 1.000 | 1 | AS_Group_CD8T_vs_Healthy_CD8T | upregulated in diseased group | MTRNR2L12 |
| 0.0001348 | -0.0778695 | 0.409 | 0.500 | 1 | AS_Group_CD8T_vs_Healthy_CD8T | No significant fold change | SSTR2 |
| 0.0002078 | 0.0159493 | 0.591 | 0.944 | 1 | AS_Group_CD8T_vs_Healthy_CD8T | No significant fold change | TNFRSF4 |
| 0.0002192 | -0.0810205 | 0.455 | 0.500 | 1 | AS_Group_CD8T_vs_Healthy_CD8T | No significant fold change | FBLN5 |
| 0.0002907 | 0.2330388 | 0.818 | 1.000 | 1 | AS_Group_CD8T_vs_Healthy_CD8T | No significant fold change | FAM102A |
| 0.0002954 | -0.0287162 | 0.364 | 0.611 | 1 | AS_Group_CD8T_vs_Healthy_CD8T | No significant fold change | CXCR3 |
| 0.0003136 | -0.0478442 | 0.545 | 0.778 | 1 | AS_Group_CD8T_vs_Healthy_CD8T | No significant fold change | SH2D1A |
| 0.0003364 | 0.0048341 | 0.545 | 0.778 | 1 | AS_Group_CD8T_vs_Healthy_CD8T | No significant fold change | RIC3 |
| 0.0003550 | -0.0310794 | 0.591 | 0.778 | 1 | AS_Group_CD8T_vs_Healthy_CD8T | No significant fold change | MFGE8 |
Gene Lists for CD14 Monocytes
data <- read.csv("Gene_Lists/DESeq2_results_AS_Group_CD14Mono.csv")
knitr::kable(head(data, 10))| p_val | avg_log2FC | pct.1 | pct.2 | p_val_adj | comparison | status | gene |
|---|---|---|---|---|---|---|---|
| 0.0000005 | 3.0793444 | 0.864 | 1 | 0.0129662 | AS_Group_CD14Mono_vs_Healthy_CD14Mono | upregulated in diseased group | MTRNR2L12 |
| 0.0000264 | -0.5630389 | 0.818 | 1 | 0.7238474 | AS_Group_CD14Mono_vs_Healthy_CD14Mono | No significant fold change | POLR2M |
| 0.0001839 | -0.6218166 | 0.864 | 1 | 1.0000000 | AS_Group_CD14Mono_vs_Healthy_CD14Mono | No significant fold change | MFN1 |
| 0.0002109 | 1.3153975 | 0.955 | 1 | 1.0000000 | AS_Group_CD14Mono_vs_Healthy_CD14Mono | upregulated in diseased group | MT-ATP8 |
| 0.0002249 | 0.1108486 | 0.864 | 1 | 1.0000000 | AS_Group_CD14Mono_vs_Healthy_CD14Mono | No significant fold change | PSMB2 |
| 0.0002389 | -0.7989884 | 0.773 | 1 | 1.0000000 | AS_Group_CD14Mono_vs_Healthy_CD14Mono | No significant fold change | CFL2 |
| 0.0002454 | 0.1862379 | 0.818 | 1 | 1.0000000 | AS_Group_CD14Mono_vs_Healthy_CD14Mono | No significant fold change | FARS2 |
| 0.0002465 | 0.2284070 | 0.818 | 1 | 1.0000000 | AS_Group_CD14Mono_vs_Healthy_CD14Mono | No significant fold change | ATG101 |
| 0.0005081 | -0.3103930 | 1.000 | 1 | 1.0000000 | AS_Group_CD14Mono_vs_Healthy_CD14Mono | No significant fold change | EIF1AX |
| 0.0005116 | 0.2612420 | 0.955 | 1 | 1.0000000 | AS_Group_CD14Mono_vs_Healthy_CD14Mono | No significant fold change | RPL26L1 |
Gene Lists for FCGR3A Monocytes
data <- read.csv("Gene_Lists/DESeq2_results_AS_Group_FCGR3AMono.csv")
knitr::kable(head(data, 10))| p_val | avg_log2FC | pct.1 | pct.2 | p_val_adj | comparison | status | gene |
|---|---|---|---|---|---|---|---|
| 0.0000003 | 3.1879492 | 0.864 | 1.000 | 0.0096059 | AS_Group_FCGR3AMono_vs_Healthy_FCGR3AMono | upregulated in diseased group | MTRNR2L12 |
| 0.0000459 | -0.5913833 | 0.727 | 1.000 | 1.0000000 | AS_Group_FCGR3AMono_vs_Healthy_FCGR3AMono | No significant fold change | SRGAP2C |
| 0.0000977 | 1.4787912 | 0.864 | 1.000 | 1.0000000 | AS_Group_FCGR3AMono_vs_Healthy_FCGR3AMono | upregulated in diseased group | MT-ATP8 |
| 0.0001116 | 0.1638615 | 0.909 | 1.000 | 1.0000000 | AS_Group_FCGR3AMono_vs_Healthy_FCGR3AMono | No significant fold change | NDUFB6 |
| 0.0001411 | -0.6162146 | 0.773 | 1.000 | 1.0000000 | AS_Group_FCGR3AMono_vs_Healthy_FCGR3AMono | No significant fold change | NPEPL1 |
| 0.0001482 | -0.3390204 | 0.955 | 1.000 | 1.0000000 | AS_Group_FCGR3AMono_vs_Healthy_FCGR3AMono | No significant fold change | BTN3A2 |
| 0.0001576 | -0.4320842 | 0.864 | 1.000 | 1.0000000 | AS_Group_FCGR3AMono_vs_Healthy_FCGR3AMono | No significant fold change | CLN3 |
| 0.0001583 | -0.4657314 | 0.682 | 1.000 | 1.0000000 | AS_Group_FCGR3AMono_vs_Healthy_FCGR3AMono | No significant fold change | CXorf23 |
| 0.0001659 | -0.5010453 | 0.773 | 1.000 | 1.0000000 | AS_Group_FCGR3AMono_vs_Healthy_FCGR3AMono | No significant fold change | CDK9 |
| 0.0002680 | 0.7405806 | 0.682 | 0.944 | 1.0000000 | AS_Group_FCGR3AMono_vs_Healthy_FCGR3AMono | No significant fold change | SLC7A5 |
Gene Lists for AS vs Healthy
data <- read.csv("Gene_Lists/DESeq2_results_AS_Group_AS_Group_vs_Healthy.csv")
knitr::kable(head(data, 10))| p_val | avg_log2FC | pct.1 | pct.2 | p_val_adj | comparison | status | gene |
|---|---|---|---|---|---|---|---|
| 0 | 3.832594 | 0.099 | 0 | 0 | AS_Group_vs_Healthy | upregulated in diseased group | VSIR |
| 0 | 4.401248 | 0.099 | 0 | 0 | AS_Group_vs_Healthy | upregulated in diseased group | ATP5MF |
| 0 | 3.136901 | 0.090 | 0 | 0 | AS_Group_vs_Healthy | upregulated in diseased group | SELENOF |
| 0 | 3.264121 | 0.090 | 0 | 0 | AS_Group_vs_Healthy | upregulated in diseased group | JPT1 |
| 0 | 3.512058 | 0.090 | 0 | 0 | AS_Group_vs_Healthy | upregulated in diseased group | SEM1 |
| 0 | 3.937238 | 0.099 | 0 | 0 | AS_Group_vs_Healthy | upregulated in diseased group | ATP5F1B |
| 0 | 3.864775 | 0.090 | 0 | 0 | AS_Group_vs_Healthy | upregulated in diseased group | ATP5PD |
| 0 | 3.940666 | 0.090 | 0 | 0 | AS_Group_vs_Healthy | upregulated in diseased group | ATP5MPL |
| 0 | 4.190652 | 0.090 | 0 | 0 | AS_Group_vs_Healthy | upregulated in diseased group | ATP5MC3 |
| 0 | 4.167637 | 0.090 | 0 | 0 | AS_Group_vs_Healthy | upregulated in diseased group | ATP5MD |
Gene Lists for All vs Healthy
data <- read.csv("Gene_Lists/Disease_Vs_Healthy_All_DEGs.csv")
knitr::kable(head(data, 10))| X | p_val | avg_log2FC | pct.1 | pct.2 | p_val_adj | Status |
|---|---|---|---|---|---|---|
| ATP6V0C | 0 | 3.764805 | 0.966 | 0.690 | 0 | Upregulated_in_diseased_group |
| PTPRCAP | 0 | 4.310562 | 0.931 | 0.517 | 0 | Upregulated_in_diseased_group |
| OVCA2 | 0 | 2.317537 | 0.931 | 0.241 | 0 | Upregulated_in_diseased_group |
| LIME1 | 0 | 3.535504 | 0.931 | 0.621 | 0 | Upregulated_in_diseased_group |
| UQCRHL | 0 | 2.538612 | 0.931 | 0.655 | 0 | Upregulated_in_diseased_group |
| NME2 | 0 | 3.205501 | 0.966 | 0.724 | 0 | Upregulated_in_diseased_group |
| MIA2 | 0 | 1.254396 | 0.759 | 0.069 | 0 | Upregulated_in_diseased_group |
| IFITM1 | 0 | 3.676931 | 0.966 | 0.862 | 0 | Upregulated_in_diseased_group |
| TEN1 | 0 | 1.056135 | 0.724 | 0.448 | 0 | Upregulated_in_diseased_group |
| SLC9A3 | 0 | 1.217472 | 0.724 | 0.000 | 0 | Upregulated_in_diseased_group |
Enrichr Results
Summary and Next Steps
This project successfully integrated scRNAseq data from AS samples and healthy controls from multiple sources. The fastq files were mapped correctly and generated count matrix, barcode, and gene list files. All samples had these three files, which were used to create Seurat object for each sample, which were merged into one large object. The combined Seurat object was batch corrected and the diseased samples were grouped. This cleaned object was used to create UMAPs to better understand the structure of the data and identify which clusters represent the cell types of interest. A DESeq2 pipeline was implemented to analyze the differential gene expression between healhty and diseased samples in both all cells as well as specific cell subsets.
However, the DESeq2 results do not reflect the expression that was expected. With low cell counts and extremely low gene expression, the subsetted cell type data is very sparse and difficult to interpret. The comparison with all cell types included showed several genes up-regulated in the diseased group related to metabolism, supporting the hypothesis that AS is driven my metabolic processes.
To correct the subsetted cell type data, I have filtered out any genes with zero counts, confirmed that the adjusted p-values are being calculated correctly, namely with the Benjamini-Hochberg procedure, and I have tried running a Wilcoxon rank sum test in place of DESeq2. I have also tried running a pipeline without pseudobulking in order to preserve cell counts, as the subsetted groups had low cell counts (less than 150). While these efforts did not create the results we were looking for, there are several other possibilities for this sparsity. In the following week, I will investigate the following:
- Confirm that pseudobulking is based on average cells per sample per cell type to maximize statistical power.
- Ensure that the sample size is sufficient for both healthy and AS groups.
- Ensure DESeq2 handles raw counts
- Re-run Wilxon rank sum test, ensuring the data is a log-normalized matrix.
- Re-evaluate my DESeq2 model and include donor-level correction.
- Run Enrichr pathway analysis on new results
These steps will confirm that the pipeline is correctly implemented so that we can have confidence in the data and can determine whether an alternative approach is needed.